{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "name": "8_Example_Ames_Housing_Dataset.ipynb", "provenance": [], "collapsed_sections": [], "toc_visible": true, "include_colab_link": true }, "kernelspec": { "name": "python3", "display_name": "Python 3" } }, "cells": [ { "cell_type": "markdown", "metadata": { "id": "view-in-github", "colab_type": "text" }, "source": [ "\"Open" ] }, { "cell_type": "markdown", "metadata": { "id": "DfQ-d2onvooo", "colab_type": "text" }, "source": [ "Mickaël Tits\n", "CETIC\n", "mickael.tits@cetic.be" ] }, { "cell_type": "markdown", "metadata": { "id": "Uvc0ah8oge9x", "colab_type": "text" }, "source": [ "# Chapitre 8 - Un exemple concret: estimation du prix d'une maison à Ames (Iowa, USA)\n", "\n", "Dans ce Chapitre, nous allons analyser un vrai Dataset de biens immobiliers: le \"Ames Housing Dataset\". A partir de ces données, nous allons développer un modèle prédictif permettant d'estimer le prix d'une maison à partir de nombreuses caractéristiques, telles que sa surface, le nombre de pièces, différents indices de qualité, etc.\n", "\n", "Plus d'informations ici: \n", "https://www.kaggle.com/c/home-data-for-ml-course/overview/description\n", "\n", "http://jse.amstat.org/v19n3/decock.pdf\n", "\n", "Détails sur les variables du dataset: https://github.com/titsitits/Python_Data_Science/blob/master/Donn%C3%A9es/data_description.txt" ] }, { "cell_type": "markdown", "metadata": { "id": "BO_L0DC-QxU9", "colab_type": "text" }, "source": [ "## Préparation/exploration du dataset" ] }, { "cell_type": "code", "metadata": { "id": "9rW_gAdoCDOm", "colab_type": "code", "colab": {} }, "source": [ "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "\n", "ames = pd.read_csv(\"https://raw.githubusercontent.com/titsitits/Python_Data_Science/master/Donn%C3%A9es/train.csv\")" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "x_SeXJLmSp2r", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 1000 }, "outputId": "0862047b-e144-4caa-f5b5-1071547046b2" }, "source": [ "ames" ], "execution_count": 371, "outputs": [ { "output_type": "execute_result", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
IdMSSubClassMSZoningLotFrontageLotAreaStreetAlleyLotShapeLandContourUtilitiesLotConfigLandSlopeNeighborhoodCondition1Condition2BldgTypeHouseStyleOverallQualOverallCondYearBuiltYearRemodAddRoofStyleRoofMatlExterior1stExterior2ndMasVnrTypeMasVnrAreaExterQualExterCondFoundationBsmtQualBsmtCondBsmtExposureBsmtFinType1BsmtFinSF1BsmtFinType2BsmtFinSF2BsmtUnfSFTotalBsmtSFHeating...CentralAirElectrical1stFlrSF2ndFlrSFLowQualFinSFGrLivAreaBsmtFullBathBsmtHalfBathFullBathHalfBathBedroomAbvGrKitchenAbvGrKitchenQualTotRmsAbvGrdFunctionalFireplacesFireplaceQuGarageTypeGarageYrBltGarageFinishGarageCarsGarageAreaGarageQualGarageCondPavedDriveWoodDeckSFOpenPorchSFEnclosedPorch3SsnPorchScreenPorchPoolAreaPoolQCFenceMiscFeatureMiscValMoSoldYrSoldSaleTypeSaleConditionSalePrice
0160RL65.08450PaveNaNRegLvlAllPubInsideGtlCollgCrNormNorm1Fam2Story7520032003GableCompShgVinylSdVinylSdBrkFace196.0GdTAPConcGdTANoGLQ706Unf0150856GasA...YSBrkr85685401710102131Gd8Typ0NaNAttchd2003.0RFn2548TATAY0610000NaNNaNNaN022008WDNormal208500
1220RL80.09600PaveNaNRegLvlAllPubFR2GtlVeenkerFeedrNorm1Fam1Story6819761976GableCompShgMetalSdMetalSdNone0.0TATACBlockGdTAGdALQ978Unf02841262GasA...YSBrkr1262001262012031TA6Typ1TAAttchd1976.0RFn2460TATAY29800000NaNNaNNaN052007WDNormal181500
2360RL68.011250PaveNaNIR1LvlAllPubInsideGtlCollgCrNormNorm1Fam2Story7520012002GableCompShgVinylSdVinylSdBrkFace162.0GdTAPConcGdTAMnGLQ486Unf0434920GasA...YSBrkr92086601786102131Gd6Typ1TAAttchd2001.0RFn2608TATAY0420000NaNNaNNaN092008WDNormal223500
3470RL60.09550PaveNaNIR1LvlAllPubCornerGtlCrawforNormNorm1Fam2Story7519151970GableCompShgWd SdngWd ShngNone0.0TATABrkTilTAGdNoALQ216Unf0540756GasA...YSBrkr96175601717101031Gd7Typ1GdDetchd1998.0Unf3642TATAY035272000NaNNaNNaN022006WDAbnorml140000
4560RL84.014260PaveNaNIR1LvlAllPubFR2GtlNoRidgeNormNorm1Fam2Story8520002000GableCompShgVinylSdVinylSdBrkFace350.0GdTAPConcGdTAAvGLQ655Unf04901145GasA...YSBrkr1145105302198102141Gd9Typ1TAAttchd2000.0RFn3836TATAY192840000NaNNaNNaN0122008WDNormal250000
5650RL85.014115PaveNaNIR1LvlAllPubInsideGtlMitchelNormNorm1Fam1.5Fin5519931995GableCompShgVinylSdVinylSdNone0.0TATAWoodGdTANoGLQ732Unf064796GasA...YSBrkr79656601362101111TA5Typ0NaNAttchd1993.0Unf2480TATAY4030032000NaNMnPrvShed700102009WDNormal143000
6720RL75.010084PaveNaNRegLvlAllPubInsideGtlSomerstNormNorm1Fam1Story8520042005GableCompShgVinylSdVinylSdStone186.0GdTAPConcExTAAvGLQ1369Unf03171686GasA...YSBrkr1694001694102031Gd7Typ1GdAttchd2004.0RFn2636TATAY255570000NaNNaNNaN082007WDNormal307000
7860RLNaN10382PaveNaNIR1LvlAllPubCornerGtlNWAmesPosNNorm1Fam2Story7619731973GableCompShgHdBoardHdBoardStone240.0TATACBlockGdTAMnALQ859BLQ322161107GasA...YSBrkr110798302090102131TA7Typ2TAAttchd1973.0RFn2484TATAY235204228000NaNNaNShed350112009WDNormal200000
8950RM51.06120PaveNaNRegLvlAllPubInsideGtlOldTownArteryNorm1Fam1.5Fin7519311950GableCompShgBrkFaceWd ShngNone0.0TATABrkTilTATANoUnf0Unf0952952GasA...YFuseF102275201774002022TA8Min12TADetchd1931.0Unf2468FaTAY900205000NaNNaNNaN042008WDAbnorml129900
910190RL50.07420PaveNaNRegLvlAllPubCornerGtlBrkSideArteryArtery2fmCon1.5Unf5619391950GableCompShgMetalSdMetalSdNone0.0TATABrkTilTATANoGLQ851Unf0140991GasA...YSBrkr1077001077101022TA5Typ2TAAttchd1939.0RFn1205GdTAY040000NaNNaNNaN012008WDNormal118000
101120RL70.011200PaveNaNRegLvlAllPubInsideGtlSawyerNormNorm1Fam1Story5519651965HipCompShgHdBoardHdBoardNone0.0TATACBlockTATANoRec906Unf01341040GasA...YSBrkr1040001040101031TA5Typ0NaNDetchd1965.0Unf1384TATAY000000NaNNaNNaN022008WDNormal129500
111260RL85.011924PaveNaNIR1LvlAllPubInsideGtlNridgHtNormNorm1Fam2Story9520052006HipCompShgWdShingWd ShngStone286.0ExTAPConcExTANoGLQ998Unf01771175GasA...YSBrkr1182114202324103041Ex11Typ2GdBuiltIn2005.0Fin3736TATAY147210000NaNNaNNaN072006NewPartial345000
121320RLNaN12968PaveNaNIR2LvlAllPubInsideGtlSawyerNormNorm1Fam1Story5619621962HipCompShgHdBoardPlywoodNone0.0TATACBlockTATANoALQ737Unf0175912GasA...YSBrkr91200912101021TA4Typ0NaNDetchd1962.0Unf1352TATAY1400001760NaNNaNNaN092008WDNormal144000
131420RL91.010652PaveNaNIR1LvlAllPubInsideGtlCollgCrNormNorm1Fam1Story7520062007GableCompShgVinylSdVinylSdStone306.0GdTAPConcGdTAAvUnf0Unf014941494GasA...YSBrkr1494001494002031Gd7Typ1GdAttchd2006.0RFn3840TATAY160330000NaNNaNNaN082007NewPartial279500
141520RLNaN10920PaveNaNIR1LvlAllPubCornerGtlNAmesNormNorm1Fam1Story6519601960HipCompShgMetalSdMetalSdBrkFace212.0TATACBlockTATANoBLQ733Unf05201253GasA...YSBrkr1253001253101121TA5Typ1FaAttchd1960.0RFn1352TATAY0213176000NaNGdWoNaN052008WDNormal157000
151645RM51.06120PaveNaNRegLvlAllPubCornerGtlBrkSideNormNorm1Fam1.5Unf7819292001GableCompShgWd SdngWd SdngNone0.0TATABrkTilTATANoUnf0Unf0832832GasA...YFuseA85400854001021TA5Typ0NaNDetchd1991.0Unf2576TATAY481120000NaNGdPrvNaN072007WDNormal132000
161720RLNaN11241PaveNaNIR1LvlAllPubCulDSacGtlNAmesNormNorm1Fam1Story6719701970GableCompShgWd SdngWd SdngBrkFace180.0TATACBlockTATANoALQ578Unf04261004GasA...YSBrkr1004001004101021TA5Typ1TAAttchd1970.0Fin2480TATAY000000NaNNaNShed70032010WDNormal149000
171890RL72.010791PaveNaNRegLvlAllPubInsideGtlSawyerNormNormDuplex1Story4519671967GableCompShgMetalSdMetalSdNone0.0TATASlabNaNNaNNaNNaN0NaN000GasA...YSBrkr1296001296002022TA6Typ0NaNCarPort1967.0Unf2516TATAY000000NaNNaNShed500102006WDNormal90000
181920RL66.013695PaveNaNRegLvlAllPubInsideGtlSawyerWRRAeNorm1Fam1Story5520042004GableCompShgVinylSdVinylSdNone0.0TATAPConcTATANoGLQ646Unf04681114GasA...YSBrkr1114001114101131Gd6Typ0NaNDetchd2004.0Unf2576TATAY01020000NaNNaNNaN062008WDNormal159000
192020RL70.07560PaveNaNRegLvlAllPubInsideGtlNAmesNormNorm1Fam1Story5619581965HipCompShgBrkFacePlywoodNone0.0TATACBlockTATANoLwQ504Unf05251029GasA...YSBrkr1339001339001031TA6Min10NaNAttchd1958.0Unf1294TATAY000000NaNMnPrvNaN052009CODAbnorml139000
202160RL101.014215PaveNaNIR1LvlAllPubCornerGtlNridgHtNormNorm1Fam2Story8520052006GableCompShgVinylSdVinylSdBrkFace380.0GdTAPConcExTAAvUnf0Unf011581158GasA...YSBrkr1158121802376003141Gd9Typ1GdBuiltIn2005.0RFn3853TATAY2401540000NaNNaNNaN0112006NewPartial325300
212245RM57.07449PaveGrvlRegBnkAllPubInsideGtlIDOTRRNormNorm1Fam1.5Unf7719301950GableCompShgWd SdngWd SdngNone0.0TATAPConcTATANoUnf0Unf0637637GasA...YFuseF1108001108001031Gd6Typ1GdAttchd1930.0Unf1280TATAN00205000NaNGdPrvNaN062007WDNormal139400
222320RL75.09742PaveNaNRegLvlAllPubInsideGtlCollgCrNormNorm1Fam1Story8520022002HipCompShgVinylSdVinylSdBrkFace281.0GdTAPConcGdTANoUnf0Unf017771777GasA...YSBrkr1795001795002031Gd7Typ1GdAttchd2002.0RFn2534TATAY1711590000NaNNaNNaN092008WDNormal230000
2324120RM44.04224PaveNaNRegLvlAllPubInsideGtlMeadowVNormNormTwnhsE1Story5719761976GableCompShgCemntBdCmentBdNone0.0TATAPConcGdTANoGLQ840Unf02001040GasA...YSBrkr1060001060101031TA6Typ1TAAttchd1976.0Unf2572TATAY1001100000NaNNaNNaN062007WDNormal129900
242520RLNaN8246PaveNaNIR1LvlAllPubInsideGtlSawyerNormNorm1Fam1Story5819682001GableCompShgPlywoodPlywoodNone0.0TAGdCBlockTATAMnRec188ALQ6682041060GasA...YSBrkr1060001060101031Gd6Typ1TAAttchd1968.0Unf1270TATAY406900000NaNMnPrvNaN052010WDNormal154000
252620RL110.014230PaveNaNRegLvlAllPubCornerGtlNridgHtNormNorm1Fam1Story8520072007GableCompShgVinylSdVinylSdStone640.0GdTAPConcGdTANoUnf0Unf015661566GasA...YSBrkr1600001600002031Gd7Typ1GdAttchd2007.0RFn3890TATAY0560000NaNNaNNaN072009WDNormal256300
262720RL60.07200PaveNaNRegLvlAllPubCornerGtlNAmesNormNorm1Fam1Story5719512000GableCompShgWd SdngWd SdngNone0.0TATACBlockTATAMnBLQ234Rec486180900GasA...YSBrkr90000900011031Gd5Typ0NaNDetchd2005.0Unf2576TATAY222320000NaNNaNNaN052010WDNormal134800
272820RL98.011478PaveNaNRegLvlAllPubInsideGtlNridgHtNormNorm1Fam1Story8520072008GableCompShgVinylSdVinylSdStone200.0GdTAPConcExTANoGLQ1218Unf04861704GasA...YSBrkr1704001704102031Gd7Typ1GdAttchd2008.0RFn3772TATAY0500000NaNNaNNaN052010WDNormal306000
282920RL47.016321PaveNaNIR1LvlAllPubCulDSacGtlNAmesNormNorm1Fam1Story5619571997GableCompShgMetalSdMetalSdNone0.0TATACBlockTATAGdBLQ1277Unf02071484GasA...YSBrkr1600001600101021TA6Typ2GdAttchd1957.0RFn1319TATAY2882580000NaNNaNNaN0122006WDNormal207500
293030RM60.06324PaveNaNIR1LvlAllPubInsideGtlBrkSideFeedrRRNn1Fam1Story4619271950GableCompShgMetalSdMetalSdNone0.0TATABrkTilTATANoUnf0Unf0520520GasA...NSBrkr52000520001011Fa4Typ0NaNDetchd1920.0Unf1240FaTAY49087000NaNNaNNaN052008WDNormal68500
......................................................................................................................................................................................................................................................
1430143160RL60.021930PaveNaNIR3LvlAllPubInsideGtlGilbertRRAnNorm1Fam2Story5520052005GableCompShgVinylSdVinylSdNone0.0GdTAPConcGdGdAvUnf0Unf0732732GasA...YSBrkr734110401838002141TA7Typ1GdBuiltIn2005.0Fin2372TATAY100400000NaNNaNNaN072006WDNormal192140
14311432120RLNaN4928PaveNaNIR1LvlAllPubInsideGtlNPkVillNormNormTwnhsE1Story6619761976GableCompShgPlywoodPlywoodNone0.0TATACBlockGdTANoLwQ958Unf00958GasA...YSBrkr95800958002021TA5Typ0NaNAttchd1976.0RFn2440TATAY0600000NaNNaNNaN0102009WDNormal143750
1432143330RL60.010800PaveGrvlRegLvlAllPubInsideGtlOldTownNormNorm1Fam1Story4619272007GableCompShgWd SdngWd SdngNone0.0TATABrkTilTATANoUnf0Unf0656656GasA...YSBrkr96800968002041TA5Typ0NaNDetchd1928.0Unf1216FaFaY000000NaNNaNNaN082007WDNormal64500
1433143460RL93.010261PaveNaNIR1LvlAllPubInsideGtlGilbertNormNorm1Fam2Story6520002000GableCompShgVinylSdVinylSdBrkFace318.0TATAPConcGdTANoUnf0Unf0936936GasA...YSBrkr96283001792102131TA8Typ1TAAttchd2000.0Fin2451TATAY000000NaNNaNNaN052008WDNormal186500
1434143520RL80.017400PaveNaNRegLowAllPubInsideModMitchelNormNorm1Fam1Story5519771977GableCompShgBrkFaceBrkFaceNone0.0TATACBlockTATANoALQ936Unf01901126GasA...YSBrkr1126001126102031TA5Typ1GdAttchd1977.0RFn2484TATAP295410000NaNNaNNaN052006WDNormal160000
1435143620RL80.08400PaveNaNRegLvlAllPubInsideGtlNAmesNormNorm1Fam1Story6919622005GableCompShgWd SdngWd SdngBrkFace237.0GdGdCBlockTATANoUnf0Unf013191319GasA...YSBrkr1537001537101131Gd7Typ1GdAttchd1962.0RFn2462TATAY0360000NaNGdPrvNaN072008CODAbnorml174000
1436143720RL60.09000PaveNaNRegLvlAllPubFR2GtlNAmesNormNorm1Fam1Story4619711971GableCompShgHdBoardHdBoardNone0.0TATAPConcTATANoALQ616Unf0248864GasA...YSBrkr86400864001031TA5Typ0NaNDetchd1974.0Unf2528TATAY000000NaNGdWoNaN052007WDNormal120500
1437143820RL96.012444PaveNaNRegLvlAllPubFR2GtlNridgHtNormNorm1Fam1Story8520082008HipCompShgVinylSdVinylSdStone426.0ExTAPConcExTAAvGLQ1336Unf05961932GasA...YSBrkr1932001932102021Ex7Typ1GdAttchd2008.0Fin3774TATAY066030400NaNNaNNaN0112008NewPartial394617
1438143920RM90.07407PaveNaNRegLvlAllPubInsideGtlOldTownArteryNorm1Fam1Story6719571996GableCompShgMetalSdMetalSdNone0.0TATACBlockTATANoGLQ600Unf0312912GasA...YFuseA1236001236101021TA6Typ0NaNAttchd1957.0Unf2923TATAY0158158000NaNMnPrvNaN042010WDNormal149700
1439144060RL80.011584PaveNaNRegLvlAllPubInsideGtlNWAmesNormNorm1FamSLvl7619791979HipCompShgHdBoardHdBoardBrkFace96.0TATACBlockTATANoGLQ315Rec110114539GasA...YSBrkr104068501725002131TA6Typ1TAAttchd1979.0RFn2550TATAY088216000NaNNaNNaN0112007WDNormal197000
1440144170RL79.011526PaveNaNIR1BnkAllPubInsideModCrawforNormNorm1Fam2.5Fin6719221994GableCompShgMetalSdMetalSdNone0.0TATABrkTilExTANoUnf0Unf0588588GasA...YSBrkr14237483842555002031TA11Min11GdDetchd1993.0Fin2672TATAY43100000NaNNaNNaN092008WDNormal191000
14411442120RMNaN4426PaveNaNRegLvlAllPubInsideGtlCollgCrNormNormTwnhsE1Story6520042004GableCompShgVinylSdVinylSdBrkFace147.0GdTAPConcGdTAAvGLQ697Unf0151848GasA...YSBrkr84800848101011Gd3Typ1TAAttchd2004.0RFn2420TATAY14900000NaNNaNNaN052008WDNormal149300
1442144360FV85.011003PaveNaNRegLvlAllPubInsideGtlSomerstNormNorm1Fam2Story10520082008GableCompShgVinylSdVinylSdStone160.0ExTAPConcExTAAvGLQ765Unf02521017GasA...YSBrkr102698102007102131Ex10Typ1ExAttchd2008.0Fin3812TATAY168520000NaNNaNNaN042009WDNormal310000
1443144430RLNaN8854PaveNaNRegLvlAllPubInsideGtlBrkSideNormNorm1Fam1.5Unf6619161950GableCompShgWd SdngWd SdngNone0.0TATABrkTilTATANoUnf0Unf0952952Grav...NFuseF95200952001021Fa4Typ1GdDetchd1916.0Unf1192FaPoP09800400NaNNaNNaN052009WDNormal121000
1444144520RL63.08500PaveNaNRegLvlAllPubFR2GtlCollgCrNormNorm1Fam1Story7520042004GableCompShgVinylSdVinylSdBrkFace106.0GdTAPConcGdTAAvUnf0Unf014221422GasA...YSBrkr1422001422002031Gd7Typ0NaNAttchd2004.0RFn2626TATAY192600000NaNNaNNaN0112007WDNormal179600
1445144685RL70.08400PaveNaNRegLvlAllPubInsideGtlSawyerNormNorm1FamSFoyer6519661966GableCompShgVinylSdVinylSdNone0.0TATACBlockTATAGdLwQ187Rec6270814GasA...YSBrkr91300913101031TA6Typ0NaNDetchd1990.0Unf1240TATAY00252000NaNNaNNaN052007WDNormal129000
1446144720RLNaN26142PaveNaNIR1LvlAllPubCulDSacGtlMitchelNormNorm1Fam1Story5719621962GableCompShgHdBoardHdBoardBrkFace189.0TATACBlockTATANoRec593Unf05951188GasA...YSBrkr1188001188001031TA6Typ0NaNAttchd1962.0Unf1312TATAP261390000NaNNaNNaN042010WDNormal157900
1447144860RL80.010000PaveNaNRegLvlAllPubInsideGtlCollgCrNormNorm1Fam2Story8519951996GableCompShgVinylSdVinylSdBrkFace438.0GdTAPConcGdTANoGLQ1079Unf01411220GasA...YSBrkr122087002090102131Gd8Typ1TAAttchd1995.0RFn2556TATAY0650000NaNNaNNaN0122007WDNormal240000
1448144950RL70.011767PaveNaNRegLvlAllPubInsideGtlEdwardsNormNorm1Fam2Story4719102000GableCompShgMetalSdHdBoardNone0.0TATACBlockFaTANoUnf0Unf0560560GasA...NSBrkr79655001346001121TA6Min20NaNDetchd1950.0Unf1384FaTAY168240000NaNGdWoNaN052007WDNormal112000
14491450180RM21.01533PaveNaNRegLvlAllPubInsideGtlMeadowVNormNormTwnhsSFoyer5719701970GableCompShgCemntBdCmentBdNone0.0TATACBlockGdTAAvGLQ553Unf077630GasA...YSBrkr63000630101011Ex3Typ0NaNNaNNaNNaN00NaNNaNY000000NaNNaNNaN082006WDAbnorml92000
1450145190RL60.09000PaveNaNRegLvlAllPubFR2GtlNAmesNormNormDuplex2Story5519741974GableCompShgVinylSdVinylSdNone0.0TATACBlockGdTANoUnf0Unf0896896GasA...YSBrkr89689601792002242TA8Typ0NaNNaNNaNNaN00NaNNaNY32450000NaNNaNNaN092009WDNormal136000
1451145220RL78.09262PaveNaNRegLvlAllPubInsideGtlSomerstNormNorm1Fam1Story8520082009GableCompShgCemntBdCmentBdStone194.0GdTAPConcGdTANoUnf0Unf015731573GasA...YSBrkr1578001578002031Ex7Typ1GdAttchd2008.0Fin3840TATAY0360000NaNNaNNaN052009NewPartial287090
14521453180RM35.03675PaveNaNRegLvlAllPubInsideGtlEdwardsNormNormTwnhsESLvl5520052005GableCompShgVinylSdVinylSdBrkFace80.0TATAPConcGdTAGdGLQ547Unf00547GasA...YSBrkr1072001072101021TA5Typ0NaNBasment2005.0Fin2525TATAY0280000NaNNaNNaN052006WDNormal145000
1453145420RL90.017217PaveNaNRegLvlAllPubInsideGtlMitchelNormNorm1Fam1Story5520062006GableCompShgVinylSdVinylSdNone0.0TATAPConcGdTANoUnf0Unf011401140GasA...YSBrkr1140001140001031TA6Typ0NaNNaNNaNNaN00NaNNaNY36560000NaNNaNNaN072006WDAbnorml84500
1454145520FV62.07500PavePaveRegLvlAllPubInsideGtlSomerstNormNorm1Fam1Story7520042005GableCompShgVinylSdVinylSdNone0.0GdTAPConcGdTANoGLQ410Unf08111221GasA...YSBrkr1221001221102021Gd6Typ0NaNAttchd2004.0RFn2400TATAY01130000NaNNaNNaN0102009WDNormal185000
1455145660RL62.07917PaveNaNRegLvlAllPubInsideGtlGilbertNormNorm1Fam2Story6519992000GableCompShgVinylSdVinylSdNone0.0TATAPConcGdTANoUnf0Unf0953953GasA...YSBrkr95369401647002131TA7Typ1TAAttchd1999.0RFn2460TATAY0400000NaNNaNNaN082007WDNormal175000
1456145720RL85.013175PaveNaNRegLvlAllPubInsideGtlNWAmesNormNorm1Fam1Story6619781988GableCompShgPlywoodPlywoodStone119.0TATACBlockGdTANoALQ790Rec1635891542GasA...YSBrkr2073002073102031TA7Min12TAAttchd1978.0Unf2500TATAY34900000NaNMnPrvNaN022010WDNormal210000
1457145870RL66.09042PaveNaNRegLvlAllPubInsideGtlCrawforNormNorm1Fam2Story7919412006GableCompShgCemntBdCmentBdNone0.0ExGdStoneTAGdNoGLQ275Unf08771152GasA...YSBrkr1188115202340002041Gd9Typ2GdAttchd1941.0RFn1252TATAY0600000NaNGdPrvShed250052010WDNormal266500
1458145920RL68.09717PaveNaNRegLvlAllPubInsideGtlNAmesNormNorm1Fam1Story5619501996HipCompShgMetalSdMetalSdNone0.0TATACBlockTATAMnGLQ49Rec102901078GasA...YFuseA1078001078101021Gd5Typ0NaNAttchd1950.0Unf1240TATAY3660112000NaNNaNNaN042010WDNormal142125
1459146020RL75.09937PaveNaNRegLvlAllPubInsideGtlEdwardsNormNorm1Fam1Story5619651965GableCompShgHdBoardHdBoardNone0.0GdTACBlockTATANoBLQ830LwQ2901361256GasA...YSBrkr1256001256101131TA6Typ0NaNAttchd1965.0Fin1276TATAY736680000NaNNaNNaN062008WDNormal147500
\n", "

1460 rows × 81 columns

\n", "
" ], "text/plain": [ " Id MSSubClass MSZoning ... SaleType SaleCondition SalePrice\n", "0 1 60 RL ... WD Normal 208500\n", "1 2 20 RL ... WD Normal 181500\n", "2 3 60 RL ... WD Normal 223500\n", "3 4 70 RL ... WD Abnorml 140000\n", "4 5 60 RL ... WD Normal 250000\n", "5 6 50 RL ... WD Normal 143000\n", "6 7 20 RL ... WD Normal 307000\n", "7 8 60 RL ... WD Normal 200000\n", "8 9 50 RM ... WD Abnorml 129900\n", "9 10 190 RL ... WD Normal 118000\n", "10 11 20 RL ... WD Normal 129500\n", "11 12 60 RL ... New Partial 345000\n", "12 13 20 RL ... WD Normal 144000\n", "13 14 20 RL ... New Partial 279500\n", "14 15 20 RL ... WD Normal 157000\n", "15 16 45 RM ... WD Normal 132000\n", "16 17 20 RL ... WD Normal 149000\n", "17 18 90 RL ... WD Normal 90000\n", "18 19 20 RL ... WD Normal 159000\n", "19 20 20 RL ... COD Abnorml 139000\n", "20 21 60 RL ... New Partial 325300\n", "21 22 45 RM ... WD Normal 139400\n", "22 23 20 RL ... WD Normal 230000\n", "23 24 120 RM ... WD Normal 129900\n", "24 25 20 RL ... WD Normal 154000\n", "25 26 20 RL ... WD Normal 256300\n", "26 27 20 RL ... WD Normal 134800\n", "27 28 20 RL ... WD Normal 306000\n", "28 29 20 RL ... WD Normal 207500\n", "29 30 30 RM ... WD Normal 68500\n", "... ... ... ... ... ... ... ...\n", "1430 1431 60 RL ... WD Normal 192140\n", "1431 1432 120 RL ... WD Normal 143750\n", "1432 1433 30 RL ... WD Normal 64500\n", "1433 1434 60 RL ... WD Normal 186500\n", "1434 1435 20 RL ... WD Normal 160000\n", "1435 1436 20 RL ... COD Abnorml 174000\n", "1436 1437 20 RL ... WD Normal 120500\n", "1437 1438 20 RL ... New Partial 394617\n", "1438 1439 20 RM ... WD Normal 149700\n", "1439 1440 60 RL ... WD Normal 197000\n", "1440 1441 70 RL ... WD Normal 191000\n", "1441 1442 120 RM ... WD Normal 149300\n", "1442 1443 60 FV ... WD Normal 310000\n", "1443 1444 30 RL ... WD Normal 121000\n", "1444 1445 20 RL ... WD Normal 179600\n", "1445 1446 85 RL ... WD Normal 129000\n", "1446 1447 20 RL ... WD Normal 157900\n", "1447 1448 60 RL ... WD Normal 240000\n", "1448 1449 50 RL ... WD Normal 112000\n", "1449 1450 180 RM ... WD Abnorml 92000\n", "1450 1451 90 RL ... WD Normal 136000\n", "1451 1452 20 RL ... New Partial 287090\n", "1452 1453 180 RM ... WD Normal 145000\n", "1453 1454 20 RL ... WD Abnorml 84500\n", "1454 1455 20 FV ... WD Normal 185000\n", "1455 1456 60 RL ... WD Normal 175000\n", "1456 1457 20 RL ... WD Normal 210000\n", "1457 1458 70 RL ... WD Normal 266500\n", "1458 1459 20 RL ... WD Normal 142125\n", "1459 1460 20 RL ... WD Normal 147500\n", "\n", "[1460 rows x 81 columns]" ] }, "metadata": { "tags": [] }, "execution_count": 371 } ] }, { "cell_type": "markdown", "metadata": { "id": "0nrcLRYFj_jN", "colab_type": "text" }, "source": [ "Commençons par explorer brièvement la qualité des données.\n", "\n", "On a 1460 observations (maisons), et 81 variables dont un Id et le prix de vente (SalePrice). 19 variables contiennent des données invalides ou manquantes." ] }, { "cell_type": "code", "metadata": { "id": "xi8-lBq2Nd0Z", "colab_type": "code", "colab": {} }, "source": [ "#Quelques fonctions utiles pour l'exploration\n", "\n", "#Par soucis de lisibilité, on affichera les series comme des dataframes d'une ligne\n", "def display_series(series): \n", " display(series.to_frame().transpose())\n", " \n", "#Corrélation entre deux colonnes\n", "def col_corr(df,col1,col2): \n", " return df[[col1,col2]].corr().values[0,1]\n", "\n", "#Pour analyser l'effet d'une variable continue sur une autre, on peut extraire deux groupes (chaque côté de la médiane), et afficher un boxplot par groupe\n", "def mediangroups_boxplot_comparison(df, group_col, comparison_col):\n", " df[\"above_median\"] = df[group_col] > df[group_col].quantile(0.5)\n", " df.boxplot(comparison_col, by = \"above_median\")\n", " df.pop(\"above_median\")\n", "\n", "#Pour analyser l'effet d'une variable continue sur une autre, on peut extraire deux groupes (chaque côté de la médiane), et calculer la moyenne par groupe\n", "def mediangroups_mean_comparison(df, group_col, comparison_col): \n", " means = df.groupby(df[group_col] > df[group_col].quantile(0.5))[comparison_col].mean()\n", " means.index = ['Below','Above']\n", " return means\n", "\n", "#Idem en séparant les groupes avec la moyenne\n", "def meangroups_mean_comparison(df, group_col, comparison_col):\n", " means = df.groupby(df[group_col] > df[group_col].mean())[comparison_col].mean()\n", " means.index = ['Below','Above']\n", " return means" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "rjnQoJAxkKi1", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 151 }, "outputId": "54603e83-02fd-41ca-c835-479804b02db9" }, "source": [ "print(ames.shape)\n", "print(list(ames.columns))\n", "\n", "#Colonnes incomplètes\n", "counts = ames.count()\n", "incomplete = counts[counts < len(ames)]\n", "display_series(incomplete.sort_values())\n", "len(incomplete.index)" ], "execution_count": 373, "outputs": [ { "output_type": "stream", "text": [ "(1460, 81)\n", "['Id', 'MSSubClass', 'MSZoning', 'LotFrontage', 'LotArea', 'Street', 'Alley', 'LotShape', 'LandContour', 'Utilities', 'LotConfig', 'LandSlope', 'Neighborhood', 'Condition1', 'Condition2', 'BldgType', 'HouseStyle', 'OverallQual', 'OverallCond', 'YearBuilt', 'YearRemodAdd', 'RoofStyle', 'RoofMatl', 'Exterior1st', 'Exterior2nd', 'MasVnrType', 'MasVnrArea', 'ExterQual', 'ExterCond', 'Foundation', 'BsmtQual', 'BsmtCond', 'BsmtExposure', 'BsmtFinType1', 'BsmtFinSF1', 'BsmtFinType2', 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF', 'Heating', 'HeatingQC', 'CentralAir', 'Electrical', '1stFlrSF', '2ndFlrSF', 'LowQualFinSF', 'GrLivArea', 'BsmtFullBath', 'BsmtHalfBath', 'FullBath', 'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'KitchenQual', 'TotRmsAbvGrd', 'Functional', 'Fireplaces', 'FireplaceQu', 'GarageType', 'GarageYrBlt', 'GarageFinish', 'GarageCars', 'GarageArea', 'GarageQual', 'GarageCond', 'PavedDrive', 'WoodDeckSF', 'OpenPorchSF', 'EnclosedPorch', '3SsnPorch', 'ScreenPorch', 'PoolArea', 'PoolQC', 'Fence', 'MiscFeature', 'MiscVal', 'MoSold', 'YrSold', 'SaleType', 'SaleCondition', 'SalePrice']\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
PoolQCMiscFeatureAlleyFenceFireplaceQuLotFrontageGarageTypeGarageYrBltGarageFinishGarageQualGarageCondBsmtExposureBsmtFinType2BsmtFinType1BsmtQualBsmtCondMasVnrAreaMasVnrTypeElectrical
07549128177012011379137913791379137914221422142314231423145214521459
\n", "
" ], "text/plain": [ " PoolQC MiscFeature Alley ... MasVnrArea MasVnrType Electrical\n", "0 7 54 91 ... 1452 1452 1459\n", "\n", "[1 rows x 19 columns]" ] }, "metadata": { "tags": [] } }, { "output_type": "execute_result", "data": { "text/plain": [ "19" ] }, "metadata": { "tags": [] }, "execution_count": 373 } ] }, { "cell_type": "code", "metadata": { "id": "3uRAS8sdiw39", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 316 }, "outputId": "472b1082-bcc0-466f-a2e2-9afc07794127" }, "source": [ "ames.hist(\"SalePrice\", bins = 30)" ], "execution_count": 374, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "array([[]],\n", " dtype=object)" ] }, "metadata": { "tags": [] }, "execution_count": 374 }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ng16SKvf/AaGB21INF5g+AAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "YNidzJlk0Ysk", "colab_type": "text" }, "source": [ "### Données manquantes" ] }, { "cell_type": "markdown", "metadata": { "id": "6na2mKJInz9N", "colab_type": "text" }, "source": [ "sur les 19 variables incomplètes, 3 sont des variables numériques (les variables de type \"float64\"), et les autres sont des variables catégorielles (les variables de type \"object\").\n", "\n", "Après vérification de la description des variables, on remarque que les variables catégorielles contiennent des \"NA\" comme catégories, et qui sont automatiquement interprétées comme des NaN par pandas. \n", "\n", "Pour les variables catégorielles, nous allons simplement remplacer les NaN par une catégorie \"Nothing\". Nous allons ensuite analyser et corriger chaque variable numérique.\n", "\n" ] }, { "cell_type": "code", "metadata": { "id": "mMH9OkrIpRWX", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 253 }, "outputId": "7f736009-4a0b-438a-d003-7b42701b6c02" }, "source": [ "#Pour vérifier le type d'une colonne: .dtypes()\n", "display_series(ames.dtypes)\n", "#Les variables catégorielles sont du type \"object\"\n", "\n", "#Vérifions les données incomplètes uniquement:\n", "incomplete_types = ames[incomplete.index].dtypes\n", "display_series(incomplete_types)\n", "\n", "#Trois variables de type \"float64\" contiennent des NaN\n", "display_series(incomplete_types[incomplete_types == \"float64\"])" ], "execution_count": 375, "outputs": [ { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
IdMSSubClassMSZoningLotFrontageLotAreaStreetAlleyLotShapeLandContourUtilitiesLotConfigLandSlopeNeighborhoodCondition1Condition2BldgTypeHouseStyleOverallQualOverallCondYearBuiltYearRemodAddRoofStyleRoofMatlExterior1stExterior2ndMasVnrTypeMasVnrAreaExterQualExterCondFoundationBsmtQualBsmtCondBsmtExposureBsmtFinType1BsmtFinSF1BsmtFinType2BsmtFinSF2BsmtUnfSFTotalBsmtSFHeating...CentralAirElectrical1stFlrSF2ndFlrSFLowQualFinSFGrLivAreaBsmtFullBathBsmtHalfBathFullBathHalfBathBedroomAbvGrKitchenAbvGrKitchenQualTotRmsAbvGrdFunctionalFireplacesFireplaceQuGarageTypeGarageYrBltGarageFinishGarageCarsGarageAreaGarageQualGarageCondPavedDriveWoodDeckSFOpenPorchSFEnclosedPorch3SsnPorchScreenPorchPoolAreaPoolQCFenceMiscFeatureMiscValMoSoldYrSoldSaleTypeSaleConditionSalePrice
0int64int64objectfloat64int64objectobjectobjectobjectobjectobjectobjectobjectobjectobjectobjectobjectint64int64int64int64objectobjectobjectobjectobjectfloat64objectobjectobjectobjectobjectobjectobjectint64objectint64int64int64object...objectobjectint64int64int64int64int64int64int64int64int64int64objectint64objectint64objectobjectfloat64objectint64int64objectobjectobjectint64int64int64int64int64int64objectobjectobjectint64int64int64objectobjectint64
\n", "

1 rows × 81 columns

\n", "
" ], "text/plain": [ " Id MSSubClass MSZoning ... SaleType SaleCondition SalePrice\n", "0 int64 int64 object ... object object int64\n", "\n", "[1 rows x 81 columns]" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
LotFrontageAlleyMasVnrTypeMasVnrAreaBsmtQualBsmtCondBsmtExposureBsmtFinType1BsmtFinType2ElectricalFireplaceQuGarageTypeGarageYrBltGarageFinishGarageQualGarageCondPoolQCFenceMiscFeature
0float64objectobjectfloat64objectobjectobjectobjectobjectobjectobjectobjectfloat64objectobjectobjectobjectobjectobject
\n", "
" ], "text/plain": [ " LotFrontage Alley MasVnrType ... PoolQC Fence MiscFeature\n", "0 float64 object object ... object object object\n", "\n", "[1 rows x 19 columns]" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
LotFrontageMasVnrAreaGarageYrBlt
0float64float64float64
\n", "
" ], "text/plain": [ " LotFrontage MasVnrArea GarageYrBlt\n", "0 float64 float64 float64" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "iC7ZwVA-XR7s", "colab_type": "text" }, "source": [ "#### Correction de variables catégorielles\n", "Pour les variables catégorielles, remplaçons les NaN par la catégorie \"Nothing\"." ] }, { "cell_type": "code", "metadata": { "id": "KgDXdSDeoitq", "colab_type": "code", "colab": {} }, "source": [ "cat_columns = ames.columns[ames.dtypes == 'object']\n", "for col in cat_columns:\n", " ames[col] = ames[col].fillna('Nothing')" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "8FfB4Ys1XVPF", "colab_type": "text" }, "source": [ "#### Exploration et correction des variables numériques\n", "Analysons la qualité de chaque variable: quelle est la proportion de données manquantes ?" ] }, { "cell_type": "code", "metadata": { "id": "l3xRRb6ep7k5", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 79 }, "outputId": "7110d85c-64d5-4520-a5af-58b86212858c" }, "source": [ "counts = ames.count()\n", "incomplete = counts[counts < len(ames)]\n", "display_series(incomplete.sort_values())" ], "execution_count": 377, "outputs": [ { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
LotFrontageGarageYrBltMasVnrArea
0120113791452
\n", "
" ], "text/plain": [ " LotFrontage GarageYrBlt MasVnrArea\n", "0 1201 1379 1452" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "8KECCDEox5Zt", "colab_type": "text" }, "source": [ "Est-ce que le fait que la variable soit manquante est en soi une information intéressante ? autrement dit, est-ce que ça a une influence sur le prix ?" ] }, { "cell_type": "code", "metadata": { "id": "CzL2lKa0r7AI", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 387 }, "outputId": "0c5a0e21-60c0-45c4-87cd-3ab5b511899e" }, "source": [ "#Mean price for NaNs and for non-NaNs\n", "display(ames.groupby(ames.LotFrontage.isna())[\"SalePrice\"].agg(['median','mean']))\n", "display(ames.groupby(ames.GarageYrBlt.isna())[\"SalePrice\"].agg(['median','mean']))\n", "display(ames.groupby(ames.MasVnrArea.isna())[\"SalePrice\"].agg(['median','mean']))" ], "execution_count": 378, "outputs": [ { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
medianmean
LotFrontage
False159500180770.480433
True172400181620.073359
\n", "
" ], "text/plain": [ " median mean\n", "LotFrontage \n", "False 159500 180770.480433\n", "True 172400 181620.073359" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
medianmean
GarageYrBlt
False167500185479.511240
True100000103317.283951
\n", "
" ], "text/plain": [ " median mean\n", "GarageYrBlt \n", "False 167500 185479.511240\n", "True 100000 103317.283951" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
medianmean
MasVnrArea
False162700180615.063361
True203287236484.250000
\n", "
" ], "text/plain": [ " median mean\n", "MasVnrArea \n", "False 162700 180615.063361\n", "True 203287 236484.250000" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "fXcg6gWatKZD", "colab_type": "text" }, "source": [ "L'absence de LotFrontage ne semble pas avoir beaucoup d'influence sur le prix. Les deux autres variables ont une influence. GarageYrBlt = NaN signifie probablement qu'il n'y a pas de garage (ce qui diminue la valeur), et MasVnrArea = NaN semble augmenter la valeur en moyenne.\n", "\n", "Pour une comparaison plus complète, nous pouvons visualiser les distributions des valeurs NaNs et non-NaN pour chacune de ces variables:" ] }, { "cell_type": "code", "metadata": { "id": "4TY52lH6cylg", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 297 }, "outputId": "a08f38d9-914e-4785-dc08-45e674eea65d" }, "source": [ "from matplotlib import pyplot as plt\n", "fig, axes = plt.subplots(1,3)\n", "\n", "for ax, feat in zip(axes, ['LotFrontage', 'GarageYrBlt','MasVnrArea']):\n", " ax.violinplot(dataset = [ames[ames[feat].isna()]['SalePrice'].values, ames[ames.notna()]['SalePrice'].values] )\n", " ax.set_xlabel(feat)\n", " ax.set_xticks([1,2])\n", " ax.set_xticklabels(['NaNs','valid'])\n", "\n", "#Permet d'éviter l'overlap entre les subplots\n", "plt.tight_layout()" ], "execution_count": 379, "outputs": [ { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlzxBOeecy6X/DyxqETvFpstqAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "8988suQTvwIX", "colab_type": "text" }, "source": [ "##### MasVnrArea" ] }, { "cell_type": "code", "metadata": { "id": "9kPorBFzt2EC", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 647 }, "outputId": "a3b60ca9-f99d-4a1b-9692-331f4fad8d42" }, "source": [ "#Analysons la relation entre \"MasVnrArea\" et le prix\n", "print(col_corr(ames,\"MasVnrArea\",\"SalePrice\"))\n", "display_series(mediangroups_mean_comparison(ames,\"MasVnrArea\",\"SalePrice\"))\n", "mediangroups_boxplot_comparison(ames,\"MasVnrArea\",\"SalePrice\")\n", "ames.plot.scatter(\"MasVnrArea\",\"SalePrice\")\n", "#Il semblerait que l'influence soit positive\n", "\n", "#Une possibilité pour prendre en compte cette information est de remplacer les NaN la moyenne de la colonne (ce qui minimisera l'influence de la variable pour ces observations particulières), et d'ajouter une colonne \"isnan_MasVnrArea\" pour garder l'information\n", "#La relation entre \"MasVnrArea\" et \"SalePrice\" ne devrait pas être trop faussée de cette manière, étant donné que le nombre de NaN est assez faible (81).\n", "ames[\"isnan_MasVnrArea\"] = ames.MasVnrArea.isna().astype(int) #.astype(int) permet de rendre la variable numérique. #.astype(str) permettrait d'en faire une variable catégorielle\n", "ames[\"MasVnrArea\"] = ames[\"MasVnrArea\"].fillna(ames[\"MasVnrArea\"].median())" ], "execution_count": 380, "outputs": [ { "output_type": "stream", "text": [ "0.4774930470957107\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BelowAbove
SalePrice157293.746835215662.741117
\n", "
" ], "text/plain": [ " Below Above\n", "SalePrice 157293.746835 215662.741117" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAEdCAYAAAAxRnE+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhQOMmZmVhQOMmZmVxf8PJ283UXOQTkEAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEKCAYAAADEovgeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "4a9-9-fqv2h6", "colab_type": "text" }, "source": [ "##### LotFrontage" ] }, { "cell_type": "code", "metadata": { "id": "Ri_eA_YQxnvA", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 647 }, "outputId": "349244b8-2bed-4969-a7cb-b157e68b1d6e" }, "source": [ "#Analysons la relation entre \"LotFrontage\" et le prix\n", "print(col_corr(ames,\"LotFrontage\",\"SalePrice\"))\n", "display_series(mediangroups_mean_comparison(ames,\"LotFrontage\",\"SalePrice\"))\n", "mediangroups_boxplot_comparison(ames,\"LotFrontage\",\"SalePrice\")\n", "ames.plot.scatter(\"LotFrontage\",\"SalePrice\", s = 1)\n", "\n", "#Ajoutons une variable isnan_LotFrontage pour pouvoir utiliser cette information dans notre modèle plus tard:\n", "ames[\"isnan_LotFrontage\"] = ames.LotFrontage.isna().astype(int)\n", "#Ensuite, étant donné le grand nombre de NaNs (~18%), il est préférable (dans une premier temps) de ne pas utiliser cette variable en remplaçant les NaN par la moyenne\n", "#Une autre possibilité serait d'entraîner un modèle de régression pour prédire ces données manquantes à partir des autres variables.\n", "out = ames.pop(\"LotFrontage\")" ], "execution_count": 381, "outputs": [ { "output_type": "stream", "text": [ "0.35179909657067854\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BelowAbove
SalePrice162409.166279207455.105
\n", "
" ], "text/plain": [ " Below Above\n", "SalePrice 162409.166279 207455.105" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAEdCAYAAAAxRnE+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZmZl4QBjZmZl8X8BKzoZ2Hex6osAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEKCAYAAADEovgeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "1Xi5TJ6r2V5p", "colab_type": "text" }, "source": [ "On ne peut pas considérer Lotfrontage = NaN comme 0, auquel cas le prix moyen serait plus faible que pour les autres maisons. Il est préférable d'omettre la variable vu la quantité de NaN (plus de 10%). Remplacer NaN par une valeur apporterait des informations erronnées, et omettre les observations réduirait significativement la taille du dataset." ] }, { "cell_type": "markdown", "metadata": { "id": "ooOgJTxIv5k8", "colab_type": "text" }, "source": [ "##### GarageYrBlt" ] }, { "cell_type": "code", "metadata": { "id": "eeIP_RARu4qW", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 413 }, "outputId": "c00aba2a-6846-46e4-9c2e-0b0e59256c09" }, "source": [ "#Concernant l'année de construction du garage, une valeur nulle n'aurait aucun sens.\n", "#On peut vérifier que la NaN correspond simplement à une absence de garage. En effet, leur surface est toujours nulle:\n", "display_series(ames[ames.GarageYrBlt.isna()].GarageArea.describe())\n", "#L'information sur l'absence de garage est donc déjà présente dans une autre variable. \n", "\n", "ames.corrwith(ames[\"GarageYrBlt\"]).sort_values(ascending = False).plot.bar(figsize = (10,4))\n", "#On remarque aussi que l'année de construction du garage est logiquement corrélée avec l'année de construction du bien.\n", "#Pour remplir les quelques NaN, on peut éventuellement \"simuler\" une date de construction du garage avec l'année de construction.\n", "ames[\"GarageYrBlt\"] = ames[\"GarageYrBlt\"].fillna(ames[\"YearBuilt\"])" ], "execution_count": 382, "outputs": [ { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
countmeanstdmin25%50%75%max
GarageArea81.00.00.00.00.00.00.00.0
\n", "
" ], "text/plain": [ " count mean std min 25% 50% 75% max\n", "GarageArea 81.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl8AAAFOCAYAAAC43Xi+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "VgsIPuREtEwY", "colab_type": "text" }, "source": [ "#### Suppression des variables inutiles" ] }, { "cell_type": "code", "metadata": { "id": "-ECUyuFItBQi", "colab_type": "code", "colab": {} }, "source": [ "out = ames.pop(\"Id\")" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "_TSUZHdqnx7U", "colab_type": "text" }, "source": [ "#### Tout en une fonction\n", "Il est intéressant de redéfinir le pipeline de correction avec une fonction, pour pouvoir le réappliquer plus tard sur de nouvelles données (le set de test par exemple)" ] }, { "cell_type": "code", "metadata": { "id": "VqIq7Ie0iM21", "colab_type": "code", "colab": {} }, "source": [ "def clean_df(df):\n", " #Fill categorical columns\n", " cat_columns = df.columns[df.dtypes == 'object']\n", " for col in cat_columns:\n", " df[col] = df[col].fillna('Nothing')\n", " \n", " #Clean MasVnrArea\n", " df[\"isnan_MasVnrArea\"] = df.MasVnrArea.isna().astype(int)\n", " df[\"MasVnrArea\"] = df[\"MasVnrArea\"].fillna(df[\"MasVnrArea\"].mean())\n", " \n", " #Clean LotFrontage\n", " df[\"isnan_LotFrontage\"] = df.LotFrontage.isna().astype(int)\n", " df = df.drop(\"LotFrontage\", axis=1)\n", " \n", " #Clean GarageYrBlt\n", " df[\"GarageYrBlt\"] = df[\"GarageYrBlt\"].fillna(df[\"YearBuilt\"])\n", " \n", " #Suppression des variables inutiles\n", " out = df.pop(\"Id\")\n", " \n", " return df" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "pOQxM9NltjAW", "colab_type": "text" }, "source": [ "## Sélection de variables prédictives" ] }, { "cell_type": "markdown", "metadata": { "id": "m308qsX1pXLP", "colab_type": "text" }, "source": [ "### Analyse de corrélation (variables continues)\n", "\n", "Remarque: les méthodes pandas.DataFrame.corr() et .corrwith() omettent automatiquement les variables non-numériques." ] }, { "cell_type": "code", "metadata": { "id": "FdC4Ay36fqTy", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 351 }, "outputId": "ad57eeae-e5f9-4347-a268-478795340f46" }, "source": [ "from matplotlib import pyplot as plt\n", "\n", "#corrélations des variables avec le prix\n", "corr_with_price = ames.corrwith(ames[\"SalePrice\"])\n", "\n", "#On trie les variables selon la valeur absolue de leur corrélation avec le prix\n", "best_features = corr_with_price.abs().sort_values(ascending=False)\n", "\n", "# On ne évidemment peut utiliser le label comme variable prédictive\n", "best_features.pop(\"SalePrice\")\n", "\n", "best_features.plot.bar(figsize=(10,4))\n", "\n", "best_features = best_features.index.to_list()" ], "execution_count": 385, "outputs": [ { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAFOCAYAAABEyFN0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEW5YOY7jOI7jVIQbVo7jOI7jOBXx/2XPXmBsesMjAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "aWA7fcF4rHOH", "colab_type": "text" }, "source": [ "### Analyse basique d'effet des variables continues" ] }, { "cell_type": "code", "metadata": { "id": "aIiC_LlqrENl", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 351 }, "outputId": "6187fb95-3581-418d-cce3-1875165d44e9" }, "source": [ "#Analyse de l'effet\n", "continuous_col = ames.columns[(ames.dtypes == 'int64') | (ames.dtypes == 'float64')]\n", "\n", "delta_mean = []\n", "\n", "for col in list(continuous_col): \n", " means = meangroups_mean_comparison(ames, col, \"SalePrice\")\n", " delta_mean.append( means[1] - means[0] )\n", "\n", "best_features_effects = pd.Series(delta_mean, continuous_col).abs().sort_values(ascending = False)\n", "best_features_effects = best_features_effects.drop(\"SalePrice\")\n", "best_features_effects.plot.bar(figsize=(10,4))\n", "\n", "best_features_effects = best_features_effects.index.to_list()" ], "execution_count": 386, "outputs": [ { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmwAAAFOCAYAAAA7GUGFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "Vi5_I8Z9yMfD", "colab_type": "text" }, "source": [ "Les résultats sont assez similaires. En réalité, les différences principales se trouvent pour les variables dont la distribution est fortement asymmétrique (e.g.: PoolArea est la plupart du temps 0, tout comme isnan_MasVnrArea)" ] }, { "cell_type": "code", "metadata": { "id": "1Fl5jS12zcMu", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 233 }, "outputId": "0ea331b1-1991-41cd-a8c1-d22f1699759d" }, "source": [ "fig, axes = plt.subplots(1,4,figsize = (15,3))\n", "ames.boxplot(\"OverallQual\", ax = axes[0])\n", "ames.boxplot(\"GrLivArea\", ax = axes[1])\n", "ames.boxplot(\"PoolArea\", ax = axes[2])\n", "ames.boxplot(\"isnan_MasVnrArea\", ax = axes[3])" ], "execution_count": 387, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "" ] }, "metadata": { "tags": [] }, "execution_count": 387 }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2oAAADGCAYAAACjHXMDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "GieK4cCJQecG", "colab_type": "text" }, "source": [ "## Préparation des sets d'entraînement et de validation" ] }, { "cell_type": "markdown", "metadata": { "id": "FZcr_oCYSmm9", "colab_type": "text" }, "source": [ "### Sélection des variables prédictives (features)" ] }, { "cell_type": "code", "metadata": { "id": "LjOypw1hSXD3", "colab_type": "code", "colab": {} }, "source": [ "#Prenons les cinq meilleures variables prédictives (en sa basant sur leur corrélation avec le prix, ou le delta de la moyenne)\n", "featurelist = best_features[:5]\n", "#featurelist = best_features_effects[:5]\n", "\n", "X = ames[featurelist]\n", "y = ames[\"SalePrice\"]" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "JCqFx1_6Sq92", "colab_type": "text" }, "source": [ "### Séparation des données d'entraînement et de validation" ] }, { "cell_type": "code", "metadata": { "id": "vDKVXwL9m5l4", "colab_type": "code", "colab": {} }, "source": [ "from sklearn.model_selection import train_test_split\n", "import numpy as np\n", "\n", "ids = ames.index\n", "\n", "trainids, valids = train_test_split(ids, test_size = 0.4, random_state = 1)\n", "\n", "#Pour avoir à chaque fois des ids différents, il suffit de ne pas fixer l'état pseudo alétoire \"random_state\":\n", "#trainids, valids = train_test_split(ids, test_size = 0.4)\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", "\n", "#Remarque: ce code est équivalent, mais le but ici est de garder les mêmes ids pour comparer plusieurs modèles (ou pour calculer la moyenne des prédictions, voir fin du notebook)\n", "#Xtrain, Xval, ytrain, yval = train_test_split(X, y, test_size = 0.4)" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "laD_Mgz0S4d1", "colab_type": "text" }, "source": [ "## Régression linéaire" ] }, { "cell_type": "code", "metadata": { "id": "9pij4mZ5l32w", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "3436b928-7298-42c4-d2ac-f8ed2074026a" }, "source": [ "from sklearn.linear_model import LinearRegression\n", "\n", "price_predictor = LinearRegression()\n", "price_predictor.fit(Xtrain, ytrain)\n" ], "execution_count": 390, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None, normalize=False)" ] }, "metadata": { "tags": [] }, "execution_count": 390 } ] }, { "cell_type": "code", "metadata": { "id": "Q4wMp6XSmJFY", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 87 }, "outputId": "cb537509-2b48-4a57-e985-cf2e0e0503ca" }, "source": [ "import numpy as np\n", "from sklearn.metrics import mean_absolute_error as MAE, mean_squared_error as MSE\n", "\n", "\n", "print(\"R2 Scores (train, val):\", price_predictor.score(Xtrain, ytrain), price_predictor.score(Xval, yval))\n", "\n", "y_pred = price_predictor.predict(Xval)\n", "\n", "print(\"biais:\", np.mean(y_pred - yval.values) ) #si il est négatif: sous-évaluation (en moyenne), si il est positif: sur-évaluation\n", "print(\"MAE:\", MAE(yval.values, y_pred) )\n", "print(\"RMSE:\", np.sqrt(MSE(yval.values, y_pred)) )\n", "\n", "#Tester en gardant ou en retirant les outliers; tester avec différentes ensembles de variables" ], "execution_count": 391, "outputs": [ { "output_type": "stream", "text": [ "R2 Scores (train, val): 0.7662346914049792 0.7477113803564559\n", "biais: 143.5871107765132\n", "MAE: 25257.404730679427\n", "RMSE: 40835.29225417492\n" ], "name": "stdout" } ] }, { "cell_type": "markdown", "metadata": { "id": "tJP043J4-Y0-", "colab_type": "text" }, "source": [ "Remarque: les résultats seront un peu différents chaque fois qu'on relancera le code, étant donné la sélection aléatoire des sets d'entraînement et de validation." ] }, { "cell_type": "markdown", "metadata": { "id": "6oTHcBp-Svi3", "colab_type": "text" }, "source": [ "## Comparaison pour différents nombres de variables prédictives" ] }, { "cell_type": "code", "metadata": { "id": "n0Rd7RnFo57A", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 313 }, "outputId": "3b938365-1ef8-4836-9cdc-e1d3da99f651" }, "source": [ "from sklearn.preprocessing import StandardScaler\n", "\n", "maes = pd.DataFrame(columns = ['train','val'])\n", "\n", "for n in range(1,len(best_features)):\n", "\n", " #Choix des variables\n", " featurelist = best_features[:n]\n", " #featurelist = best_features_effects[:n]\n", " \n", " #Commencer par les plus mauvaises\n", " #featurelist = best_features[::-1][:n]\n", " #featurelist = best_features_effects[::-1][:n]\n", "\n", " X = ames[featurelist]\n", " y = ames[\"SalePrice\"] \n", " \n", " Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", " #scaler = StandardScaler()\n", " #Xtrain = scaler.fit_transform(Xtrain)\n", " #Xval = scaler.transform(Xval)\n", " price_predictor.fit(Xtrain, ytrain)\n", " trainpred = price_predictor.predict(Xtrain)\n", " valpred = price_predictor.predict(Xval)\n", " \n", " maes.loc[n] = MAE(ytrain, trainpred), MAE(yval, valpred)\n", "\n", "maes.plot()\n", "plt.title(\"Performances (MAE) pour différents nombres de variables prédictives\")\n", "plt.xlabel(\"Nombre de variables prédictives\")\n", "plt.ylabel(\"Maen Absolute Error\")" ], "execution_count": 392, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "Text(0, 0.5, 'Maen Absolute Error')" ] }, "metadata": { "tags": [] }, "execution_count": 392 }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbwAAAEXCAYAAADWa/HyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncRynKnCF5ziO41QF/w+UhPNdgLL9qAAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "code", "metadata": { "id": "W0xYzhPQ1lRN", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "01462152-80e5-4bba-f64f-89b5f4a3a26f" }, "source": [ "maes['val'].min(), maes['val'].idxmin()" ], "execution_count": 393, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(21187.228717528884, 28)" ] }, "metadata": { "tags": [] }, "execution_count": 393 } ] }, { "cell_type": "markdown", "metadata": { "id": "7b-C0Mb14Hzo", "colab_type": "text" }, "source": [ "* On peut constater que les 10-12 variables prédictives les moins corrélées à SalePrice sont (logiquement) assez inefficaces pour prédire le prix.\n", "* Les meilleurs résultats en validation sont obtenus avec les 25-30 variables les plus corrélées au prix. (cela peut varier selon la sélection du set d'entraînement)\n", "\n" ] }, { "cell_type": "code", "metadata": { "id": "dswF-30GVnAN", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 439 }, "outputId": "4fe9ae6d-1c0b-416a-e2c5-74b71cbf402b" }, "source": [ "#Meilleur résultat en validation\n", "featurelist = best_features[:28]\n", "\n", "X = ames[featurelist]\n", "y = ames[\"SalePrice\"]\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", "\n", "price_predictor.fit(Xtrain, ytrain)\n", "trainpred = price_predictor.predict(Xtrain)\n", "valpred = price_predictor.predict(Xval)\n", "\n", "def display_results(labels, predictions, title = \"Model results\", figsize = (10,6), **kwargs):\n", " \n", " print(\"Mean average error: \", MAE(labels, predictions)) \n", " plt.scatter(labels, predictions, s=2, label = title, **kwargs)\n", " plt.gcf().set_size_inches(figsize)\n", " plt.title(title)\n", " plt.xlabel(\"Labels\")\n", " plt.ylabel(\"Prédictions\")\n", " prev_legend = plt.legend()\n", " plt.legend()\n", " xmin, xmax = plt.gca().get_xlim()\n", " plt.plot([xmin,xmax],[xmin,xmax], 'k')\n", "\n", "print(MAE(ytrain, trainpred), MAE(yval, valpred))\n", "\n", "display_results(yval, valpred, \"Régression linéaire - validation set\")" ], "execution_count": 394, "outputs": [ { "output_type": "stream", "text": [ "20316.58741840932 21187.228717528884\n", "Mean average error: 21187.228717528884\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAn0AAAGDCAYAAABeA6Y5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPiGEEEIIJyBBnxBCCCGEE/j/zBLqSJ/WJjwAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "uugb_WZ_hh8D", "colab_type": "text" }, "source": [ "On constate que le nuage de doit semble former une légère courbe. Cela semble indiquer qu'il existe des relations non-linéaires, telles que des relations polynomiales entre les variables et le prix. La régression linéaire se base en effet sur l'hypothèse très simpliste que les relations seraient linéaires. Si on revisualise les nuages de points plus haut montrant l'interactions entre différentes variables (comme 'OverallQual') et le prix, on constate que la relation et plutôt quadratique que linéaire.\n", "\n", "Une variation de la régression linéaire, appelée régression polynomiale, permet de facilement prendre en compte de potentielles interactions non-linéaires entre les variables, en créant de nouvelles variables à partir de polynômes des variables d'origine. Bien sûr, les possibilités ne se liminent pas aux polynômes." ] }, { "cell_type": "markdown", "metadata": { "id": "TAI-or-klHpt", "colab_type": "text" }, "source": [ "##Régression polynomiale" ] }, { "cell_type": "code", "metadata": { "id": "EgiED5-Ug39a", "colab_type": "code", "colab": {} }, "source": [ "def add_polynomials(X):\n", " #(remove some unwanted warnings)\n", " prev = pd.options.mode.chained_assignment\n", " pd.options.mode.chained_assignment = None\n", " for col in X.columns:\n", " #prise en compte des carrés des variables\n", " X[col+\"_square\"] = X[col]**2\n", " #prise en compte des racines carrées\n", " X[col+\"_sqrt\"] = X[col]**(1/2)\n", " \n", " pd.options.mode.chained_assignment = prev\n", " \n", " return X" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "0Cp45O_3hEqJ", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 439 }, "outputId": "f329c598-9b0e-48de-dbaf-b16fbf0fd1f4" }, "source": [ "featurelist = best_features[:28]\n", "\n", "X = ames[featurelist]\n", "X = add_polynomials(X)\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", "price_predictor.fit(Xtrain, ytrain)\n", "trainpred, valpred = price_predictor.predict(Xtrain), price_predictor.predict(Xval)\n", "\n", "print(\"MAE (train,val):\", MAE(trainpred, ytrain), MAE(valpred, yval))\n", "display_results(yval, valpred, \"Régression polynomiale (sqrt et square) - validation set\")" ], "execution_count": 396, "outputs": [ { "output_type": "stream", "text": [ "MAE (train,val): 18935.991226609982 20230.204028521497\n", "Mean average error: 20230.204028521497\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAoUAAAGDCAYAAABdmGpUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngoJCEREREUFBoYiIiIigoFBEREREgP8PO2c+0vv3avcAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "OO_nP6e-ru99", "colab_type": "text" }, "source": [ "Les prédiction semblent suivre une ligne droite. On remarque par contre que certaines prédictions parfois sont très mauvaises, et parfois même négatives (et peuvent aller jusqu'à -5.000.000 selon l'état pseudo-aléatoire utilisé).\n", "\n", "Un simple bornage des prédictions entre les prix minimum et maximum du set d'entraînement permet de limiter ces erreurs exceptionnelles.\n", "\n", "Néanmoins, selon les objectifs, il convient de définir une métrique d'évaluation appropriée: si quelques rares erreurs très grandes sont tolérables, ou si il est préférable d'avoir des prédictions globalement un peu moins précises mais sans gros écarts.\n", "Dans tous les cas, on peut facilement améliorer les résultats en contraignant les prédictions entre des valeurs limites." ] }, { "cell_type": "code", "metadata": { "id": "nKU4XFodjXLX", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 129 }, "outputId": "0b44210b-343b-47a6-ad38-7784cdf558f6" }, "source": [ "lowest_pred_index = yval.index[valpred.argmin()]\n", "display_series(ames.loc[lowest_pred_index])" ], "execution_count": 397, "outputs": [ { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MSSubClassMSZoningLotAreaStreetAlleyLotShapeLandContourUtilitiesLotConfigLandSlopeNeighborhoodCondition1Condition2BldgTypeHouseStyleOverallQualOverallCondYearBuiltYearRemodAddRoofStyleRoofMatlExterior1stExterior2ndMasVnrTypeMasVnrAreaExterQualExterCondFoundationBsmtQualBsmtCondBsmtExposureBsmtFinType1BsmtFinSF1BsmtFinType2BsmtFinSF2BsmtUnfSFTotalBsmtSFHeatingHeatingQCCentralAir...1stFlrSF2ndFlrSFLowQualFinSFGrLivAreaBsmtFullBathBsmtHalfBathFullBathHalfBathBedroomAbvGrKitchenAbvGrKitchenQualTotRmsAbvGrdFunctionalFireplacesFireplaceQuGarageTypeGarageYrBltGarageFinishGarageCarsGarageAreaGarageQualGarageCondPavedDriveWoodDeckSFOpenPorchSFEnclosedPorch3SsnPorchScreenPorchPoolAreaPoolQCFenceMiscFeatureMiscValMoSoldYrSoldSaleTypeSaleConditionSalePriceisnan_MasVnrAreaisnan_LotFrontage
129860RL63887PaveNothingIR3BnkAllPubCornerGtlEdwardsFeedrNorm1Fam2Story10520082008HipClyTileStuccoStuccoStone796ExTAPConcExTAGdGLQ5644Unf04666110GasAExY...469295005642202131Ex12Typ3GdAttchd2008Fin21418TATAY214292000480GdNothingNothing012008NewPartial16000000
\n", "

1 rows × 81 columns

\n", "
" ], "text/plain": [ " MSSubClass MSZoning LotArea ... SalePrice isnan_MasVnrArea isnan_LotFrontage\n", "1298 60 RL 63887 ... 160000 0 0\n", "\n", "[1 rows x 81 columns]" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "code", "metadata": { "id": "Jt96KbO9sFfg", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 439 }, "outputId": "f7900434-d556-4c10-a7e2-5629669cbcac" }, "source": [ "trainpred = trainpred.clip(ytrain.min(), ytrain.max())\n", "valpred = valpred.clip(ytrain.min(), ytrain.max())\n", "\n", "print(\"MAE (train,val):\", MAE(trainpred, ytrain), MAE(valpred, yval))\n", "display_results(yval, valpred, \"Régression polynomiale (sqrt et square) corrigée - validation set\")" ], "execution_count": 398, "outputs": [ { "output_type": "stream", "text": [ "MAE (train,val): 18891.879496570047 19434.25335313196\n", "Mean average error: 19434.25335313196\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAn0AAAGDCAYAAABeA6Y5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nERERiQAK+kREREQigII+ERERkQjw/wGu/OunzATpvwAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "XZoJLuVkVIRx", "colab_type": "text" }, "source": [ "## Comparaison pour différents nombres d'exemples (taille du training set)" ] }, { "cell_type": "code", "metadata": { "id": "q74RICIET8YG", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 769 }, "outputId": "71ceb7eb-2b2a-4cee-cf4c-4ed3ba6e544d" }, "source": [ "maes2 = pd.DataFrame(columns = ['train','val']) \n", "\n", "plot_limit = 0\n", "leg = []\n", "\n", "for nsamples in range(10,800,10):\n", " \n", " \n", " trainset, valset = train_test_split(ames, test_size = 660, random_state=1)\n", " \n", " #On extrait un sample sur le training set\n", " trainset_sample = trainset.sample(nsamples, random_state=1)\n", " \n", " Xtrain = trainset_sample[featurelist]\n", " ytrain = trainset_sample[\"SalePrice\"]\n", " \n", " Xval = valset[featurelist]\n", " yval = valset[\"SalePrice\"]\n", " \n", " price_predictor.fit(Xtrain, ytrain)\n", " \n", " trainpred = price_predictor.predict(Xtrain)\n", " valpred = price_predictor.predict(Xval) \n", " trainpred = trainpred.clip(ytrain.min(), ytrain.max())\n", " valpred = valpred.clip(ytrain.min(), ytrain.max()) \n", " maes2.loc[nsamples] = MAE(ytrain, trainpred), MAE(yval, valpred)\n", " if nsamples in [10,50,100,790]: \n", " display_results(yval, valpred, \"Résultats pour nsamples = \" + str(nsamples))\n", "\n", "maes2.plot()\n", "plt.title(\"Performances (MAE) pour différentes tailles de set d'entraînement\")\n", "plt.xlabel(\"Nombre d'observations d'entraînement\")\n", "plt.ylabel(\"Mean Absolute Error\")" ], "execution_count": 399, "outputs": [ { "output_type": "stream", "text": [ "Mean average error: 62075.304060146256\n", "Mean average error: 66187.58959638455\n", "Mean average error: 29063.136947200164\n", "Mean average error: 21553.222971613843\n" ], "name": "stdout" }, { "output_type": "execute_result", "data": { "text/plain": [ "Text(0, 0.5, 'Mean Absolute Error')" ] }, "metadata": { "tags": [] }, "execution_count": 399 }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAGDCAYAAAC8371AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3Xd4VGX6//H3k0Ag9BKQ3pQiUhJp\nKkXqAmtbEUFxrVgiqLisroCrrhXL/nTRhc2K7AqubVG/rF1BQBSVmhBAiojUUAyQBAgh7fn9cc6E\nyWSSTEKGScLndV1zZebMKfeZieT2KfdjrLWIiIiISOUTFuoARERERCQ4lOiJiIiIVFJK9EREREQq\nKSV6IiIiIpWUEj0RERGRSkqJnoiIiEglpURPRAowxsw0xjxdBucZaIzZUxYxSdkyxtxijPk21HGI\nSHAp0RM5SxhjdhhjThhjjhlj9htjXjfG1PKz351AprX24SDFMLQE+79ujHmqrOOQ8sUYE+f+Xnoe\nJ40xR73eP98Ys9gYk2qM2WaMudrn+CHGmM3GmHRjzBJjTOszfxci5ZMSPZGzyxXW2lpANBADTPXd\nwVr7qrX2D2c8sgrKGFMl1DFUdNbaWGttLc8DeBuYD3mf7/+Aj4EGwJ3Af4wxHdz3o4APgEfc91cD\n7575uxApn5ToiZyFrLX7gS9wEj4AjDHVjDF/NcbsMsYccFtZIt33oowxHxtjUowxh40x3xhjwtz3\nrDHmPK/z+G2FM8a8AbQCPnJbbf7kbp/vtjCmGmOWGWMucLffCdwA/Mnd/yN3+0PGmL3GmKPGmC3G\nmCH+7tGNI84Ys9Dd92vvlh5jzCXGmFXudVcZYy7xei9fy6Mx5i/GmP+4z9u49zzeGLMLWOzn2gON\nMXuMMX80xhw0xuwzxtzq9f5vjTE/unHtNcY84G6v737OvxpjjrjPW3gdt9QY85Qx5jvPZ2KMaWiM\nedMYk+beRxuv/a0x5j5jzHZjTLIx5gXP9+Yn5k7uZ3XY/VzHFBdvMBhjagLXAHPdTZ2AZsBL1toc\na+1iYDlwo/v+KGCjtXa+tTYD+AvQ3RjTKVgxilQkSvREzkJu8jAS2Oa1+VmgA07ydx7QHHjUfe+P\nwB6gEXAOMA0o0fqJ1tobgV24rYrW2ufdtz4D2gONgbXAm+7+r7rPn3f3v8IY0xG4B+hlra0NDAd2\nFHHZG4AngSggwXNuY0wD4BPgZaAh8CLwiTGmYQlu6VLgfDcGf5oAdXE+x/HATGNMffe9OcBd7j10\n4VSyGAb8G2iNkxSfAP7uc97rcJKc5sC5wPfuMQ2ATcBjPvtfDfQELgSuAm7zDdRNrhYCb+F8D9cB\ns4wxnYuJ1/c8/dz/GSjs0a+Qz8rbNcCvwLIi9jFuHAAXAOs8b1hrjwM/u9tFznpK9ETOLgvcsU+7\ngYO4SYExxuB0if3BWnvYWnsUeAbnDz5AFtAUaG2tzbLWfmPLaKFsa+2/rLVHrbUnOdUaU7eQ3XOA\nakBnY0xVa+0Oa+3PRZz+E2vtMvfcDwMXG2NaApcBP1lr37DWZltr3wY2A1eUIPS/WGuPW2tPFPJ+\nFvCE+3l9ChwDOnq919kYU8dae8RauxbAWnvIWvu+tTbd/Q6exkkovf3bWvuztTYVJ0n+2Vq7yFqb\njdPdGeOz/3Pud7oL+BtwvZ9YLwd2WGv/7X4e8cD7wLVFxevLWvuttbZeEY9AJn/cDMzz+v3agvO7\n+qAxpqox5jfuZ1LDfb8WkOpzjlSgdgDXEqn0lOiJnF1+57bKDMTpEotytzfC+cO5xtP6Anzubgd4\nAaf170u3G3BKWQRjjAk3xjxrjPnZGJPGqda5KH/7W2u3AffjJIQHjTHvGGOaFXGJ3V7HHgMO43QD\nNgN2+uy7E6eVLFC7i3n/kJt8eaTjJCXgtFr9FtjpdilfDGCMqWGM+acxZqf7eSwD6hljwr3Oc8Dr\n+Qk/r30n2HjHuRPn3n21Bvp4t77htIY2KSresmaMaYXzuznPs81amwX8Dic534/TuvxfnBZmcBLo\nOj6nqgMcRUSU6Imcjay1XwOvA391NyXjJAkXeLW+1HUHxuO2uP3RWtsOuBKY7DU2Lp1TrStwKjnw\ne2mf1+NwuhOH4nRztnG3m0L2x1r7lrW2H05yYoHnirheS88T48wwbgAkuQ/fmZmtgL3u8+MUf0+l\nbtG01q6y1l6F0026ACdxASeJ6Qj0sdbWAQZ4wi/ttfD6DHDuMcnPPruBr31a32pZa+8uJt58jDH9\nTf7Zs76P/sXEeiOw3Fq73XujtTbRWnuptbahtXY40A5Y6b69EejuFUNNnC7tjcVcS+SsoERP5Oz1\nN2CYMaa7tTYXmA28ZIxpDGCMaW6MGe4+v9wYc57bxZuK04Wa654nARjnts6NoGBXo7cDOH+kPWoD\nJ4FDOInVM0Xtb4zpaIwZbIypBmTgJKe5FO637rixCJyxej9Ya3cDnwIdjDHjjDFVjDFjgc44Mzs9\n93Sd21XYExhdxDVKxBgTYYy5wRhT122tSvO6h9ruPaW44wh9x9uVxoPuJI+WwCT8z0j9GOfzuNG9\n56rGmF7GKWtSVLz5uF36tYp4fFNMrDfh/A9IPsaYbsaY6m6L5wM4wwg8+/0f0MUYc40xpjrOuNJE\na+3m4j8akcpPiZ7IWcpa+ytOF5lnwsVDON2zP7jdhos4Naasvfv6GM7g/1nW2iXue5NwxrZ5uvsW\nFHHZ6cCf3e7BB9zr78RpSfsR+MFn/zk4Y8NSjDELcMbnPYvTArkfp4WpQIkYL2/hJEuHgR7A7917\nP4QzLu2POEnmn4DLrbXJ7nGP4LQKHQEed89Tlm4EdrifcyzO5wZO8h2Jc38/4HSfn67/AWtwktdP\ncD7TfNzxgL/BGZOZhPPZPofzeRcVb5lxu4Nb4JZV8XEjsA9nrN4QYJg77tLze3wNznjGI0AfTo0t\nFTnrmTIaTy0iUq4YY14H9lhr/xzqWELFGGOB9u7YRhE5C6lFT0RERKSSUqInIiIiUkmp61ZERESk\nklKLnoiIiEglpURPREREpJKqEuoAyouoqCjbpk2bUIchIiIiUqw1a9YkW2sbFbefEj1XmzZtWL16\ndajDEBERESmWMcZ3GUe/1HUrIiIiUkkp0RMRERGppJToiYiIiFRSSvREREREKikleiIiIiKVlBI9\nERERkUpKiZ6IiIhIJaVET0RERKSSUqInIiIiUkkp0RMRERGppJToiYiIiFRSSvREREREykB2djbv\nvPMO1tpQh5JHiZ6IiIjIafrqq6+Ijo7m+uuvZ9GiRaEOJ48SPREREZFS2rFjB9dccw1Dhw4lPT2d\nBQsWMHTo0FCHladKqAMQERERqWjS09N57rnneP755wkLC+Ppp59m8uTJVK9ePdSh5aNET0RERCRA\n1lrmz5/PAw88wO7du7n++ut5/vnnadGiRahD80tdtyIiIiIBSExMZNCgQYwdO5aGDRuybNky3nrr\nrXKb5IESPREREZEiHTp0iIkTJxITE8OGDRuIi4tj9erV9O/fP9ShFUtdtyIiIiJ+ZGdn8+qrr/LI\nI4+QmprKxIkTefzxx6lfv36oQwuYEj0RERERH19//TX33XcfiYmJDB48mBkzZtClS5dQh1Vi6roV\nERERce3atYuxY8cycOBAUlNTee+991i0aFGFTPJALXoiIiIinDhxghdeeIFnn30WgMcff5wHH3yQ\nyMjIEEd2epToiYiIyFnLWssHH3zAH//4R3bu3MmYMWN44YUXaNWqVahDKxPquhUREZGz0oYNGxg6\ndCijR4+mTp06LFmyhHfffbfSJHmgRE9ERETOMocOHeK+++4jOjqa+Ph4Zs6cydq1axk4cGCoQytz\n6roVERGRs0JmZia33XYbb7/9NgB33XUXTz75JA0bNgxxZMGjRE9EREQqvVmzZvHAAw9w4sQJ6tat\ny5tvvslll10W6rCCTomeiIiIVFqrVq3i2muvZefOnYSHhzNp0iRefPFFwsLOjtFrSvRERESk0klJ\nSWHMmDEsXLgQgP79+/P+++/TqFGjEEd2Zp0d6ayIiIicFXJzc5k2bRqNGjVi4cKFNGvWjK+//ppl\ny5addUkeKNETERGRSuLjjz+mcePGTJ8+nfDwcJ577jn27t3LgAEDQh1ayKjrVkRERCq0Xbt2cfXV\nV7N27VoARo0axRtvvEGNGjVCHFnoqUVPREREKqTs7GxuvfVW2rRpw9q1a+nUqRM//vgj77//vpI8\nlxI9ERERqXBeffVV6taty+uvv06tWrX4z3/+w6ZNmzj//PNDHVq5oq5bERERqTDi4+MZPXo027dv\nJywsjAkTJvDKK6+cNeVSSkqJnoiIiJR7aWlpXHfddXz22WcAXHLJJbz//vs0adIkxJGVb0p/RURE\npFz7y1/+QlRUFJ999hlNmjThq6++Yvny5UryAqBET0RERMql2bNnExkZyeOPP44xhqeffpp9+/Yx\nePDgUIdWYajrVkRERMqVTZs2MWjQIA4cOABAv379+Pzzz6lZs2aII6t41KInIiIi5UJmZiaXXnop\nnTt35sCBA9SpU4clS5bwzTffKMkrpaAmesaYPxhjNhpjNhhj3jbGVDfGtDXGrDDGbDPGvGuMiXD3\nrea+3ua+38brPFPd7VuMMcO9to9wt20zxkzx2u73GiIiIlI+Pfroo0RGRrJs2TLCwsJ49NFHSU1N\nZeDAgaEOrUILWqJnjGkO3Af0tNZ2AcKB64DngJestecBR4Dx7iHjgSPu9pfc/TDGdHaPuwAYAcwy\nxoQbY8KBmcBIoDNwvbsvRVxDREREypElS5ZQp04dnnzySXJzcxk4cCAnTpzg8ccfD3VolUKwu26r\nAJHGmCpADWAfMBh4z31/LvA79/lV7mvc94cYY4y7/R1r7Ulr7S/ANqC3+9hmrd1urc0E3gGuco8p\n7BoiIiJSDhw+fJjOnTszePBgjh49StOmTdmyZQtLliwhIkIdcWUlaImetXYv8FdgF06ClwqsAVKs\ntdnubnuA5u7z5sBu99hsd/+G3tt9jilse8MiriEiIiIhdtNNN9GwYUM2bdpEREQEc+bMISkpiQ4d\nOoQ6tEonmF239XFa49oCzYCaOF2v5YYx5k5jzGpjzOpff/011OGIiIhUavPmzaNatWq88cYbAFx/\n/fWcPHmS2267LcSRVV7BLK8yFPjFWvsrgDHmA6AvUM8YU8VtcWsB7HX33wu0BPa4Xb11gUNe2z28\nj/G3/VAR18jHWvsq8CpAz5497endroiIiPizbds2Lr30UpKSkgDo1KkTy5cvp0GDBiGOrPIL5hi9\nXcBFxpga7ri5IcCPwBJgtLvPzcD/3Ocfuq9x319srbXu9uvcWbltgfbASmAV0N6dYRuBM2HjQ/eY\nwq4hIiIiZ0hOTg5Dhgyhffv2JCUlUbt2bb788ks2bdqkJO8MCeYYvRU4EyLWAuvda70KPARMNsZs\nwxlPN8c9ZA7Q0N0+GZjinmcj8F+cJPFzYKK1NsdtrbsH+ALYBPzX3ZciriEiIiJnwBNPPEFERASL\nFy8mLCyMKVOmkJaWxrBhw0Id2lnFOA1g0rNnT7t69epQhyEiIlKhffTRR1x77bWcPHkScFa1+Oqr\nrzSTtowZY9ZYa3sWt5+WQBMREZHTlpycTNeuXdm/fz8ADRo04Ouvv6ZLly4hjuzspiXQRERE5LSM\nHDmSRo0asX//fowx/OUvf+HQoUNK8soBteiJiIhIqcyYMYPJkyeTm5sLON20S5YsoUoVpRflhb4J\nERERKZGNGzdy8cUXc/ToUQAaNmzImjVraN26dYgjE1/quhUREZGAZGdn0717d7p06cLRo0cJDw9n\n7ty5JCcnK8krp5ToiYiISLEmTZpE1apVSUxMBGDs2LFkZ2dz0003hTgyKYq6bkVERKRQn3/+OVdd\ndRWZmZkAtG3blsTERGrVqhXiyCQQatETKcfWLf4vn119CesW/zfUoYjIWSYlJYUWLVowcuRIMjMz\nqV69OsuWLWP79u1K8ioQJXoi5VjSK3+jzaYjJL3yt1CHIiJnkSuuuIL69euzd+9ejDFMmzaNEydO\n0L9//1CHJiWkRE+kHGt27/3sOL8+ze69P9ShiMhZIC4ujvDwcD7++GMAevfuTWZmJk8//XSII5PS\n0hg9kXKs++AxdB88JtRhiEglt2XLFvr06UNqaioA9erVY9WqVZx33nkhjkxOl1r0REREzlLZ2dn0\n6NGDTp06kZqaSnh4OLNnz+bIkSNK8ioJJXoiIiJnoQcffJCIiAjWrl0LwKhRo8jOzub2228PcWRS\nltR1KyIichZZvHgxl112GRkZGQC0atWK9evXU6dOnRBHJsGgRE9EROQscOzYMS644AJ27doFQLVq\n1fj0008ZPHhwiCOTYFLXrYiISCU3evRoateuza5duzDGMHnyZDIyMpTknQXUoiciIlJJzZkzh7vu\nuoucnBwALrzwQlasWEGVKvrzf7bQNy0iIlLJ/PLLL/To0YMjR44AULduXVasWEHHjh1DHJmcaeq6\nFRERqSSys7Pp06cP7dq148iRI4SFhTFz5kxSUlKU5J2llOiJiIhUAg8//DARERGsXLkScJYxy8nJ\nYcKECSGOTEJJXbciUqSEgwnErYsjtnss0Y2jQx2OiPj49ttvGTZsWF65lObNm7Nhwwbq1asX4sik\nPFCLnogUKW5dHMuTlhO3Li7UoYiIl2PHjnHuuefSv39/MjIyiIiI4LPPPmPPnj1K8iSPEj0RKVJs\n91j6NutLbPfYUIciIq7rr7+e2rVrs337dgDuvfdeTp48yYgRI0IcmZQ36roVkSJFN44mbpha80TK\ngzfeeINbb701r1xK165dWbt2rcqlSKH0myEiIlLO7d69m5iYGA4dOgRA7dq1+f7777ngggtCHJmU\nd+q6FRERKaeys7Pp168frVq14tChQ4SFhfHSSy+RlpamJE8CokRPRESkHHr88ceJiIhg+fLlAAwf\nPpycnBzuv//+EEcmFYm6bkVERMqRH374gSFDhpCeng5AkyZNWL9+PVFRUSGOTCoiteiJiIiUAxkZ\nGXTo0IGLL76Y9PR0qlatyoIFC9i3b5+SPCk1JXoiIiIhdssttxAZGclPP/0EQGxsLJmZmVx11VUh\njkwqOnXdioiIhMh///tfbrjhBrKzswHo3Lkz69atU7kUKTP6TRIRETnD9u/fT7du3fj1118BqFWr\nFl9//TUXXnhhiCOTykZdtyIiImfQoEGDaNq0Kb/++ivGGKZPn87Ro0eV5ElQKNETERE5A5599lnC\nwsJYunQp4CR8ubm5TJkyJbSBSaWmrlsREZEgWrt2LQMGDOD48eMANG7cmHXr1tGkSZMQRyZnA7Xo\niYiIBEFGRgbnn38+PXr04Pjx41SpUoX58+dz4MCBcpvkJRxMIHZhLAkHE0IdipQRJXoiIiJl7I47\n7iAyMpLNmzcDTvmUrKwsRo8eHeLIiha3Lo7lScuJWxcX6lCkjKjrVkREpIz83//9H2PHjiUrKwuA\nDh06sG7dOqpXrx7iyAIT2z0230+p+JToiYiInKbk5GS6du3K/v37AahRowZLly6lV69eIY6sZKIb\nRxM3TK15lYm6bkVERIpR1Ni14cOH06hRI/bv348xhieeeILjx49XuCRPKie16ImIiBTDM3YNyGvx\nevHFF3nwwQfJzc0FoH///ixevFirWki5ot9GERGRYniPXUtMTKRv374cO3YMgKioKOLj42nRokUo\nQxTxS4meiIhIMaIbR/P3QX8nJiaGDRs2AFClShXmzp3LuHHjQhydSOE0Rk9ERKQY9957L1WrVs1L\n8m644QaysrKU5Em5pxY9ERGRQnz88cdcc801ZGZmAtCuXTs2bNhAZGRkiCMTCYwSPRERER+HDx+m\na9euJCUlARAZGcmXX35Jv379QhyZSMmo61ZERMTLZZddRsOGDUlKSsIYw7Rp00hPT1eSJxWSWvRE\nRESAmTNnct999+WVS7nooov45ptvVC5FKrSgtugZY+oZY94zxmw2xmwyxlxsjGlgjFlojPnJ/Vnf\n3dcYY142xmwzxiQaYy70Os/N7v4/GWNu9trewxiz3j3mZWOMcbf7vYaIiIivLVu2ULduXe655x5y\nc3Np0KABv/zyC99//72SPKnwgt11OwP43FrbCegObAKmAF9Za9sDX7mvAUYC7d3HncA/wEnagMeA\nPkBv4DGvxO0fwB1ex41wtxd2DREREQCys7OJiYmhU6dOpKWlER4ezr///W8OHTpEmzZtQh2eSJkI\nWqJnjKkLDADmAFhrM621KcBVwFx3t7nA79znVwHzrOMHoJ4xpikwHFhorT1srT0CLARGuO/Vsdb+\nYK21wDyfc/m7hoiICJMnT6Zq1aokJDhLmo0ePZrs7GxuueWW0AYmUsaC2SbdFvgV+LcxpjuwBpgE\nnGOt3efusx84x33eHNjtdfwed1tR2/f42U4R1xARkbPYwoULueKKKzh58iQArVu3ZsOGDdSqVSvE\nkYkERzC7bqsAFwL/sNbGAMfx6UJ1W+JsEGMo8hrGmDuNMauNMat//fXXYIYhIiIhlJqaSqtWrfjN\nb37DyZMnqV69OkuWLGHHjh1K8qRSC2aitwfYY61d4b5+DyfxO+B2u+L+POi+vxdo6XV8C3dbUdtb\n+NlOEdfIx1r7qrW2p7W2Z6NGjUp1kyIiUr5dffXV1KtXj927d2OM4cEHH+TEiRMMHDgw1KGJBF3Q\nEj1r7X5gtzGmo7tpCPAj8CHgmTl7M/A/9/mHwE3u7NuLgFS3+/UL4DfGmPruJIzfAF+476UZYy5y\nZ9ve5HMuf9cQEZGzxOzZs6lSpQoLFiwAoGfPnmRmZvL888+HODKRMyfY88bvBd40xkQA24FbcZLL\n/xpjxgM7gTHuvp8CvwW2AenuvlhrDxtjngRWufs9Ya097D6fALwORAKfuQ+AZwu5hoiIVHLbtm2j\nV69epKSkAFCvXj1WrlxJ+/btQxyZyJlnnCFs0rNnT7t69epQhyEiIqWUnZ3NxRdfjOff8vDwcP7+\n978TGxsb4shEyp4xZo21tmdx+2kJNBERqfAeeughIiIi8pK83/3ud2RnZyvJk7OeSn6LiEiFtWzZ\nMoYPH05GRgYALVq0YP369dSrVy/EkYmUD2rRExGRCufYsWO0bduWSy+9lIyMDKpVq8aXX37J7t27\nleSJeFGiJyIiFcrYsWOpXbs2O3bsAGDSpElkZGQwbNiw0AYmUg6p61ZERCqEuXPnMn78eHJycgCI\njo5m1apVVKmiP2UihdF/HSIiUq7t2LGDHj16cPiwU1mrTp06fP/993Tu3DnEkYmUf+q6FQmG3Svh\njVGw+nWYPcR57F4Z2LGrX4dnWsDMPiTPu5msvzQged7Nzvm8z+G5hp9tO76cSeL0IWxetajw/f0d\n72f7ivn/jyN/acHGufef2l7Esamzr+Chl2bzu5nLWbPzSOGfjZ/PY/OqRSROH8KOL2f6v1YRx5YX\nnnvw/uzX7DzCTXNW5H0e/vYpT3zjDZXs7GwuueQS2rZty+HDhwkLC+Pll18mNTX1jCd55f07EymM\nWvREgmHps/DzV5AUDycOn9p24wfFH/vV45B5FH7dTINfNxMGNNi+4NT7nnN4ruFnW4Ptq2hj00hc\nNB16DfW/v7/j/ezXYeNL1OcotX6ZB+Sc2q+QY+vuXcbInBTezZrCjEVbmTe+j//PxvdYIHPRdLqd\nXE3a91vBphW8VmHXLUc89+D92c9YtJVlPyUDMG98H7/7lCe+8YbCo48+ylNPPYWn1uvIkSP59NNP\nQxIL+P9eRSoCJXoiwTBwivPz/Csh/o3824oz5DH48s9QtzmHa3ei7vaPSG13BVHmaP5zeJ772Xb4\nnCHsWPMBEUOnFr6/v+P9bN96wR/osPElktqO5oKwXf5j8HqdmpHFZ8euJDqiHpOGdih4f4VdF4gY\nOpXERdOp02MUdQ58Vfi1Av0sQ8BzD96fvedz8Pz0t0954hvvmfT9998zZMgQTpw4AUCzZs1Yt24d\nUVFRZzwWb+X9OxMpjFbGcGllDDnT1uw8woxFW5k0tAM9WtcPdThyliivv3cZGRl06dKFn3/+GYCI\niAjef/99Lr/88hBHJlI+aWUMkXLO0z02Y9HWkh9cAcaqSfl0Wr93QfL73/+eyMjIvCTv7rvv5uTJ\nk0ryRMqAum5FQuS0useKGOcmUpRQdsv6evvtt7npppvIzs4GoEuXLsTHx6tcikgZUtetS123clp2\nr3SSr4FToGXvync9kTK0Z88eYmJiSE52JnzUqlWL5cuX061btxBHJlJxqOtW5EzytLAtffbMXK9l\nb6clT0leuRCqciTlpQxKSVx66aW0bNmS5ORkwsLC+Otf/8rRo0eV5IkEiRI9kbIwcAqcO6T42aBl\nPbZOY/XKhVCNeyuP4+0K89RTTxEWFsayZcsAGDp0KDk5Ofzxj38McWQilZsGQoiUBU8LW3HKemyd\nxuqVC6Ea91aextsVZtWqVQwaNIjjx48DcM4555CQkECTJk1CHJnI2UGJnsiZVEQNuXJxPimVHq3r\nh6SwcKiuG4iMjAy6d+/O1q1Oa2PVqlV55513GDVqVIgjEzm7qOtW5EzyM7butMZZaayelEO33347\nkZGReUne+PHjyczMVJInEgJK9ERCrCKNswpYGY4drIgTDspMBRuD+f777xMREcGcOXMA6NixIydO\nnOC1114LcWQiZy913YqEWEUYZ1ViZTh2sDysuxoyFWQM5v79++nevTsHDx4EoGbNmixbtowLL7ww\nxJFJial0U6WjFj2REPOMs/K3HFWFbc0KdBZyACYN7cCA9lGVKxEOVBl+jsEydOhQmjZtysGDBzHG\n8NRTT3Hs2DEleRXVmS4VJUGngskuFUyW8uimOStY9lMyA9pHVarWrPK63qoE7oUXXuChhx7C8zdk\nwIABfP311yGOSk6bWvQqjEALJqvrVqQcq5TdugS/O/asSCRD9Ac5MTGRvn37cuzYMQAaNWpEYmKi\nyqVUFoGWipIKQ123IuVYUd26FVmwu2NDOcEl4WACsQtjSTiYENwLneEutuzsbLp06UL37t05duwY\nVapU4e233+bgwYNK8kTKMSWBMcanAAAgAElEQVR6IsEQwtmSFWFcX7AT2FCO64tbF8fypOXErYsL\n7oXO4Pi9u+++m6pVq7Jx40YAbrzxRrKysrjuuuuCfm0ROT1K9ESCIYQDmitluZYSCmVLaGz3WPo2\n60ts99jgXqhlbyfJW/psyf6HogT/E/LRRx8RERFBXJyTtLZv354TJ04wb9680kYt5Vx6fDy7br+D\n9Pj4UIciZUSJnkgwlKa1xfMHePXrp9UaGEhrVkVo9SutUN9bdONo4obFEd04OvgXK83/UARwTHJy\nMs2aNePKK68kKyuLyMhIvvvuO7Zu3Ur16tXLIHApr5JnzuL4t9+SPHNWqEORMqLJGCLBUJoBzZ4/\nwEnxcOKws60Ug6IDWRarMtemq8z3VkBplsAr5piRI0fy+eefA2CM4ZFHHuHxxx8/nSilAomaOCHf\nT6n4lOiJlBeeP7znXwmbPgzq2KvKOpsXKve9FVCK/6FIqBZBXJPGxFaLwLvN8eWXX+YPf/gDubm5\nAPRt34ClX3xElbaXlGHAUt7ViImh1WuzQx2GlCHV0XOpjp6InCkJBxOIWxdHbPfY0+riLc15YhfG\nsjxpOX2b9SVuWBwbN27kkksuIS0tDYAGDRqw5uEetElb4Qw/UKkNkXIp0Dp6AY3RM8ZMMsbUMY45\nxpi1xpjfnH6YIlJehXqsW2VWVjNzS3Mez2SR2y+4ne7du9OlSxfS0tIIDw/n9ddf59ChQ7S59qly\nvyKHiAQm0MkYt1lr04DfAPWBGwGtjyJSzpRlcuYZ63b73FVK9spYaWbm+vtuS3Oe6MbRVPu4Gr2a\n9yIxMRGAMWPGkJ2dzc033+zs5OkS1soIIhVeoImecX/+FnjDWrvRa5uInGmFlMgoq9Iqa3YeIe1E\nFrWrVeFIelalLtUSipbL0szMfWbRZ6zKeI5nFn1W6vN88cUXVK9enZdffhmAtm3bcvToUd59992S\n3YCIVBiBJnprjDFf4iR6XxhjagO5wQtLpBIqrn5ZSYosF1Iiw1NaZVq3o6dVomXGoq0k7Enl3EY1\ng1N4OIQFpX1VlLqDEY2+okqtn4ho9FWJj01JSaFFixaMGDGCkydPUr16db7++mu2b99OrVq1ghCt\niJQXgSZ644EpQC9rbToQAdwatKhEKqPi6peVpCZaIXX6PKVVOm2edVoFmz0J4yNXXBCcwsMhLCjt\nK5SraJTEQ+2G0JcaPNRuiP8dCkmer7rqKurXr8/evXsxxjB16lROnDjBgAEDCr3WGVvGrQJRIWGp\nqAIqr2KtzTXGHAA6G2NUkkWkNIqreVaSmmjFldUoTX01L4HU4jstpxlfWQr6vZaR6Pj3iPtlM4S9\nB13GFZxx60meAW78gLi4OCZOnJhXLqVXr1589913VKlS/D/hnkkeAHHDgryUWwXhKSQMqPyIVCgB\nJW3GmOeAscCPQI672QLLghSXSOVTXHJWmiLLZ+JcxViz8wgzFm1l0tAOgbf8ncH4Kg2f5LhAMuZu\n39ryenrXq0dqaioA9erVY9WqVZx33nkBX8ozuSPoy7hVICokLBVVQHX0jDFbgG7W2pPBDyk0VEdP\npHRumrOCZT8lM6B9VIVoGQu2UiW+peDbopednU2fPn1Yu3YtAOHh4cyaNYs777wzaDGISOiUaR09\nYDtQ9fRCEqnkiplgUFnr0pXFGLfK9Nn4mx0bDN4zbh988EEiIiLykryrr76a7OxsJXkiEvASaOlA\ngjHmKyCvVc9ae19QohKpiHzGSPmqbGuwerdcne79FPXZlNUqEmdKRKOvqHLkJyLqfwWMC+q1Fi9e\nzGWXXUZGRgYArVq1Yt26ddSrVy+o1xWRiiPQRO9D9yEihSlmgkFlW4O1LBPXoj6bijYx4KGL7s1L\nTAu1e6XzPwYDp5SqKPGxY8do36kz+/fuBqBatWp8+umnDB48uLRhi0glFeis27nGmAjA86/wFmtt\nVvDCEql8erSuz6ShHfjkkwWcV/UD6o54pEKvPFCSxLW4cWv+Zr56jrms1w1AxZkY4OlSLVIxrb9F\nGT16NO+//37e607DrmfTl2+VNEwROUsEOut2IDAX2IGzIkZLY8zN1lrNuhXxCOCP94xFW7lt37+o\nG57o7F+BZ56WpCxJaVr/Th0Txbzx5b8lr0RKUV5mzpw53HXXXeTkOIUPOnXpRsw9s/jjiM7BiFBE\nKolAu27/H/Aba+0WAGNMB+BtoEewAhOpcAL44+206N1GTNUPqFsOasidKaXptq6wXd2BdMuWoLzM\njh07uPDCCzlyxJmoUrduXVasWEHHjh3z7VfRxjKKyJkRaHmVRGttt+K2VWQqryJnq7IqB3KmyoqE\nQonu7Y1RTsvuuUNOq8U2OzubmD4xbFi7AYCwsDBefvllJk6c6Hf/2IWxLE9aTt9mfSvEWEYROT1l\nXV5ltTHmNWPMQPcxG1BWJBIs/kq1BGl92LJa67XE5ynifspbuZUS3Vshy9OVxMMPP0xERERekteq\ndytycnIKTfLAGcPYt1nfCjOWUUTOjEC7bu8GJgKecirfALOCEpGI+B/v52/bac7ehLLrIi3xeYoY\n01jeStEUe2++30MpW/K+/fZbhg0bllcupXHTxoycOZL7+95f7LEBTQIRkbNOQF23ZwN13Uq54i+B\n87etjLoJg6XILk+f+/HeF+CTTxYwqaLMTj7N7+HYsWN0796d7du3AxAREcH//vc/RowYUdaRkh4f\nT/LMWURNnECNmJgyP7+InBll0nVrjPmv+3O9MSbR9xFgIOHGmHhjzMfu67bGmBXGmG3GmHfdsi0Y\nY6q5r7e577fxOsdUd/sWY8xwr+0j3G3bjDFTvLb7vYZIheFpFfJOcPxtK4NuwmAqssvT637W7DzC\n7XNX5e0bHrmTpJpx/JL8g5MMlnen8T2MGzeO2rVr5yV599xzDydPngxKkgeQPHMWx7/9luSZ6pQR\nORsU13U7yf15+WlcYxKwCajjvn4OeMla+44xJg4YD/zD/XnEWnueMeY6d7+xxpjOwHXABUAzYJE7\n6xdgJjAM2AOsMsZ8aK39sYhriFQup9FNeCYE2p07Y9FWjqRnUb9GVSYN7UDcuqksJx2atSOunCax\n+ZTie3jzzTe5+eab88qldO3albVr11KlSqAjakonauKEfD9FpHIrskXPWrvPfTrBWrvT+wEU+6+E\nMaYFcBnwmvvaAIOB99xd5gK/c59f5b7GfX+Iu/9VwDvW2pPW2l+AbUBv97HNWrvdWpsJvANcVcw1\nRCq98jSRwVNrr7iZqp71cl+7uRc9Wtc/NbFg5D/Lf7dtCe3Zs4eoqCh+//vfk5OTQ+3atdmwYQOJ\niYlBT/IAasTE0Oq12eq2FTlLBDrrdpifbSMDOO5vwJ+AXPd1QyDFWpvtvt4DNHefNwd2A7jvp7r7\n5233Oaaw7UVdQ6TSK6y7tNAE8DRn85ZFYumbEHomFlSmenDZ2dn079+fli1bcujQIcLCwnjppZdI\nS0vjggsuCHV4IlJJFTdG725jzHqgk8/4vF+A9cUcezlw0Fq7pgzjLVPGmDuNMauNMat//fXXUIcj\nUiY8rWO+3aUFEkBPgvf5VGciQSnHwpVVeZaQCSTR9d6nFInxE088QUREBN9++y0Aw4cPJycnh/vv\nL342rYjI6Siun+At4DNgOuA9UOaotfZwMcf2Ba40xvwWqI4zRm8GUM8YU8VtcWsB7HX33wu0BPYY\nY6oAdYFDXts9vI/xt/1QEdfIx1r7KvAqOLNui7kfkdALoJxKYUuTFRgv5ylv0rznaU3o8JxvWrej\nTgLkia0MSr+UiaLi2L0S3hoLJ5x/zhKGP+p/dQnvUjAQ8Dq1K1euZNCgQaSnpwPQpEkT1q9fT1RU\n1GnflohIIIobo5dqrd2Bk6Ad9hqfl22MKbK4lbV2qrW2hbW2Dc5kisXW2huAJcBod7ebgf+5zz90\nX+O+v9g6tV8+BK5zZ+W2BdoDK4FVQHt3hm2Ee40P3WMKu4ZI+VLS1iFPwlGK1rcC4+U8M0VHTC84\nm7cU5+20eVb+2E4j1qKUuKu4qDiWPuskeZENYOAU4tbFsTxpOXHrfOrRec+qDWCGbUZGBh06dKBP\nnz6kp6dTtWpVFixYwL59+5TkiciZZa0t9gHE49bcc1+HAWsDOdbdfyDwsfu8HU6itg2YD1Rzt1d3\nX29z32/ndfzDwM/AFmCk1/bfAlvd9x722u73GkU9evToYUXOhNU7DtsbX/vBrt5x2Np5V1v7WB3n\nZyB2rXD23bUiOPGczn6+sQUhVmutvfG1H2zrhz62N772Q2AHFBWHz3vxB+LtXV/eZeMPxJc6vt9f\ncYUF8h6xsbGlPldxjq9da3eOv90eX7s2aNcQkfIJWG0DyMECXes2wVob7bNNa92KlMJNc1aw7Kdk\nBrSPYt5vzJnr3vTqwlyT2z6vOLFnjN2A9lFFrkKRL+4grVaRcDDBf9epl/K6pu78+fMZN24c2dnO\nPLD2devyY3KyM5PWp/s4kPsMxK7b7+D4t99Ss18/Wr02u6xuRUWVRSqAsl7rdrsx5j5jTFX3MQnY\nfnohilQ+gXQr5pss4a8IcmkEMlnAqwvTewJFYZM3iow7AKWZjZvXdfrZXYV2ZwdasqXQOMp4zeD9\n+/fTuHFjxowZQ3Z2NjUjI/nkyqtIWLLkVLkUn+7jQruISyA9Pp6ctDSqd+tW5jXxVFRZpPIItGhT\nLPAy8Gec7oivgDuDFZRIRRXIGq2FTZY4LYFMFvCMKRs4hUm57QHyWsUCiaekcef7LAJsuYztHgtJ\n8cQmbXf2L20xaK8WtBmLbP7vpIg1dos7l2/sgwYNYunSpQAYA888NIEp02cWPMf5V0JSvPPTc59e\nPwPh28qWPHMWGYmJ/NymC4cbtKFHwGcqnooqi1QeASV61tqDOJMdRKQIga4EUea8krgC2zy8Vm/o\nQeGJaFmZNLQD5538kUm5cfB5Oux1h0YUkVxFN47m/s5Pk5k0nc2dJtDJd4dAZ/J6krmMVF6hBjNa\njeKyoZc47/l8VsV2o/pJDJ977jmmTp3qGRPMoPMbsnhMFpy7r+DxAJs+dCZ9bPoQet6SVyewJDyt\nbACtXptN1MQJJO5J4e+tBtJ40VbiLowos+5WT1FlEan4ikz0jDF/stY+b4x5BaclLx9r7X1Bi0yk\nAuoR9hPzIp6DsCk4i7cUo6xKkPguwVUOlkXr0bo+PWp/BD8vK1EJl2cSa7MsdTIDEmszr5fPm0uf\nJWHPN8Qt3kvsyH/mJWaecXsjujTl8w37mNbNTRIzUqm7dxmPnlsVWt/qnMPns/J0owKnki/v78Ur\nMVy7di0DBgzg+PHjADRu3Jh169bRJGvXqf398ZeIl5BvK1uNmBhypg6n4eZZXHNeNZJnrsiXCIqI\nQPEtepvcn5qlIBKIYroFC0wkKGk3IpSf+nSB8E5wvGMt4h6KbBUdOIW4xXuddXDXxeUlZp5u4vV7\nUzmSngVEMW/8B/mv4/L9Dvx2o/p8LyfHvE1MTAybNjn/JFapUoW3336b0aM9VZyaFP39lXJN4nyx\n+mll+2zHbBKrZ1F7x2yGTXQ+C3W3ioi3IhM9a+1H7s+5Re0nIq5iWm4KjOErYv9CZ5eWJjkshTKZ\n3VpYglPEPRQ5FrBlb2J7TII1M4htOiBvsycp9LToeV4nVIsgrkljYqtF4OmU9f0O/Hajen0vd955\nJ7Nnn0qwbr75Zl5//XXn/AHOni3tLNvixnzmfRY9JlGji7pbRaSg4rpuP8JPl62HtfbKMo9IJFRK\n2VKWPyEquuWmQGtVES09hf6RL4NuwEAEMrGk1E7jHqLj3yPul80Q9h50GQecSg4TDiawLO3fhEfG\nAvX9dssGNI6yZW8W1L6JMef2IysrC4AOHTqwbt06NqdtJnZhLLFNBxC3ZobTuuh1fn/8dg8HoLhY\no7uMI879DERE/Cmu6/av7s9RQBPgP+7r64EDwQpKJCRK2VJWkoSoJDNXC/0jX8puwJIK5sQSfy1t\nASfaXkmiJ8me1u0onTbPIq5eBMsPb4SMFOJSMomNcbpWvbtli/sOkpOT6dq1K/v37wegRo0aLF26\nlF69nAGDcd+4SZtndnCzdsXOni3NLNtAYhURKU6gBZNX+xbl87etIlPBZCmbFr3QFPANJIayKtJb\nFmIXxrI8aTl9m/U91cL1xign0T53SMCJrKeI84d1X6TbydUktLuEuKatid23k+jt3xV7rrzPpOkA\nouPfY/jMrXy5YiMAxhgee+wxHnvssSKP8f19KctiwypcLCKFKeuCyTWNMe28Tt4WqFna4ETOmJIU\nxy1l8eKSFvAtUiDx+tnnyY9/ZNlPyTz58Y+FHubpPpzwybMlKmJcEunx8ey6/Q7S4+MLvOdduDi2\neyx9m/XN38LltYZsoMWWPUWcI4ZOhXOHED3oceKGxRE96PG8cyUcTCB2YSwJBxMKHJ/3mTz9EOE3\n/V9ekte/QwMyMzPzkjzvc3TYa5n2bg4dss53Yl76LPN/eJ7+7/Rn/pb5ZVpsWIWLReR0BVow+Q/A\nUmPMdsAArYG7ghaVSFkprDu2vM5cDaT72N8+bsv88Ywsbpqzwm/LXmz3WBL3pLB/1wBmLNoalC5B\n31pv3ny7uAuMVfPqkn7viTcZtngef9sXwR/vfLLQFsh8XZu9hp5KggdOcc61eyVxn92Vbxydd8vm\nkIgh/Ovuf5F1whmH17BWDT7p2ZauU6adWtWC/GPspr2bw/Fvv2XjoY00vbou0du/4+Xcn0khl5fj\nX+aLiS8DRc9+DbSlToWLReR0BdSiZ639HGgPTALuAzpaa78IZmAiZcKrlSgfnyWpQsq7hW7gFKfm\nXEZq4a16A6eQ2nwATxy9Iq/F65ErLmBA+yhqVq+at7SZr+jG0bwy+B/0a9EjKOPu1uw8wj9aDSSn\nRx+OjBvG9R+O55o5b+XF6L2EWoFWNp9Wyhu2LKTn/l2MXL6tZMuEud9r4n+mOtdd+iyxSdvpS428\n1sO4dXF8vetrBvcezJiBY8g6kUV4eDhPz3yaN4Z2p86+HJLnL8l3Wu8WyKiJE9hxfn1m90wjrn5d\nOHcI93W8gXrV6nFfzH15xYa9Ezjf+w20pc7fuURESiKgRM8YUwN4ELjHWrsOaGWMuTyokYmUhcK6\nYwtLAMtSoN3G3klny95Qva6zikRhSWjL3twb9mf+tatxXkLnadka07Ml9WtUZUSXpgUOe2vFLm6f\nu4oRXZrmtfaVZj3awsxYtJW30+vx2ODfcM/BV9hwZCU/ZrxXIMYerU/Nhr138b1O8uOTeLd7YBI5\nvbux9erofN27RXXDAjBwCtuqdiI7/QiffLKAhJjRxDVrR2yPSXmtginvpLD59s0cce953LhxpK1a\nRczCj/m8dSrr29bm0DU35jutpwRLdONoasTE0PSfs2jcux+xfZyWw2sv+hPfXPcN13a81m9Yvmvb\nRk2cQM1+/dRSJyJBV2jXrZvILbXWHgP+DawBLnbf3gvMBz4OeoQiwXAmZq4GOovXt9RIAKVHCpsR\n+/mGfRxJz+LzDfsY16dVvvde+GIzR9KzeOGLzXnvlWUJFU8sWY0+IOVICrWq1KVZ9dF+Ww9ju8ey\n8dBGUk6mELcujjife64RE0OXee/SBfJ1x8Zt/lfRZUpa9qZR48act3cZ51b9gIf2NXe6bfct44qd\ndRk1ahSZmZkA1I6qwqd/aEa/aW+y6/Y7aLPpCH0zajPloluISdpDo4WxhU5ciT6ZSdz+g9DJOVdx\nQwF8Z91qiTEROVOKGqO3HYgDfg+ca60da4y5HsBam26MMWciQJEKK9Bacb5JZwBJaIGyG26i8fvm\nt7B+b3W/LXoPDu/EC19s5sHhp1aQLbSEiu+KEgGMZ8yrZbfhZ+LWbCC2+wSiC6nx5nQjv5I3Vo7G\n0QXv2RNDRmreOrmxwx91fvqUKck3o3jEI7D0WeoOnEJstQjqrT/Av+6exz9P/BOAyMhIhk6N5pdW\nR/lPzTb049QYuPbX3Ei/pGpkNXqV5UlOS6zfhNJN4tP3pJP8c0uizt1NjZPfO+/5+e5Ks7atiEhZ\nKDTRs9b+aIyZ6r7MNMZE4hZPNsacC5w8A/GJVFzBbjX0TsbcxKPJnlSOpE/226I3rk+rAtsKrdPm\n3RoJBVsmC2nBSjiYQNyaGU59uaw4Yt1VG/ISPq/jolv2Ljr58cTgtU6u2bCVq2YlYu7dCoNPtbQV\nKEjsxvnwZZfx6aefAs4ssqnTpvH000+zbvF/SXrlbzS71+mi3drcEDc2nNjudZg3MpqEg9VOJaF+\npDceTfLbu8iJCCdj87ccTz2frAt7EtHJXWNXRKScKG4JtN3u08eAz4GWxpg3gb7ALcENTaTiC2p9\nOzcRSs3IYkbWKCY1zyIiehIDEmuf9mSLzZ0mkLkn1UlcmtRxNnq3THolgmsGzMm7x9lb45yu0mbt\nIPcYy8li45rneAVnRQvv1rlik2A/6+Qm3Xc/bTYdYccrf6P74DF5u8Z2jyUtM420zDTmb5nPjFdm\n8N0/vsPmOrORo+s34OsvPqeOW/S4/lsLidh0hJpvLYTBYwokit4tcP5myCbPX8Lx7cep3q0WNfv1\nY37keTT+7Ht+zDrA9F4FbyVYtRZVZ09EilPsZAy3i3YzzuoYtwBvAz2ttUuDGplIReQzAcMzBs7f\nLFgP34H6AXMnlMzIGsW/djXm3rA/06nX0DKp6fdMYm2uTJ3MM4m1SagWQWyTxiRUiyhwbQZOyXeP\nsU0H0JcajGxzB9VOXkNda0ghl7g1M061CvqbBOP1ueVNuKgWATd+QHpy1bzafM3uvZ8d59en2b33\n5zs8unE0dSLqsGr9Kq6LuY7lM5djcy11qlbly9Zt+N+oUXlJHkDKRZ1Ir1mFlIuc9je/df1c/mbI\neiZTnDN1Cq1em83ww5voeXALN2xZ6PfzfG/upwybO5335n5akq+hWKqzJyLFKbaOnrXWGmM+tdZ2\nBT45AzGJVEhrdh6h6ltT6XbyVIuVp2VtRJemRda38/4ZMLdr+LKdR9jmthb5i6k0LUneY/fi1k0t\nOAHCq1t60tBT5VOilz1H3C+bSUx6m/+lTuaqNnAy8kNi214FYc0LH+fn1UIY16Qxy5OWk5aZRp2I\nOtzzxhHCVyYC0P212bSv357kmbNIr98+rxUrOzubpZOXsm3DNgCMgds7N6HhiOZ0TIkpMLt17yfv\n0+Z4Njs+eZ/zegymwcxZvDhxAjXcFlXvljJ/texqxMRw+JmJPLMujtiDsbS+ciQHf95CkytH+v08\nb9iykPCDW4jZshC4ocxa4lRnT8paeVjpR8pWoAWT1xpjellrVwU1GpEKbMairRxLu5y/1IFubouV\nZwycZ6kuKDi7tciB+rtXwufuUNkR0/0mSUWth+o9q3byFdUC7iL2Pmd4ZNGJaN6+u1c6XbPNexIR\nPZUeiQc5FraAe5N2EB22OqCZx+mNR3PPG/+jSb8ubAWWJy2nSb8u3BZxqhSJd1Hmw89M5Ka7b2L9\nB+vzTtW2f1teP+dcGq3fw9G9TWn1dsHZrc3uvZ8dr/yN5pddw+67J5CbkgKcKvLsW/jZ3wzZBQue\nY8D/JbDg6lRuW1mL3JQUji5cRP2xYwvs2+6BSXmJnb/zl5Zm70pZK8uZ+FI+BLoEWh/gB2PMz8aY\nRGPMemNMYjADE6loJg3tQHTLerSOquH3PU+xYKBkNfb2ri66rp4fnu7Py3plMqB9FNO6HXVWiChF\nF7F3Dbl8fO/BE2v1unTqNZRGrZaxMiyDuGbt8hK5wurgJVSLYFyD6nw162+Er0zktpW1+FPry+lL\nDX53yRX5igZ7uk3XXRhDjxY98pK81q1bc/MHN1NzfE0WD4uiZr9+dPnTE37vqfvgMYz8v++o98Nm\nclNSCKtXL1+rmHedu8JiHv1tLjG/OD99u4J9+RY+PnTNjfzcpkuBen0ioVbg3yqp8AJN9IYD7YDB\nwBXA5e5PEXH1aF2fR2t/RN29y4pPygJcmWNzpwlsq9qJY1HRbO40IeDixp5xf0sOvMm88X3otHkW\nsUnb6Z1bnV93DSj8HCVZG9jnHhJiRjO+dQcmHe+Rt55t16iupEWdmze+zxPXggXP5VsT97kfXmF9\n8nr+3TudHefXJ2riBKLj3yPul83OJA4v2eeeS78vv+DKiRPJzcolvAq8ems3Pr6oC41+XEfX2m05\n/6IJ/Pni29m5aTW7ftuT9C/e8nsLnoSu5T9mUSMqK+/evROzwsZQtvvDFGr260e7P0xh7yfvU+N4\nNns/eb/4zw14Kaka90TfwktJ1QLaX+RMKdO1u6VcKDLRM8ZUN8bcj7Mqxghgr7V2p+dxRiIUqUgK\nWXGjwKSMAFfmeCaxNkOPPsqEyOd5JrF2sRM7PAadcwO1cy9g0Dk35F0vukV/Wp+YwJqf6hZ+jhIs\nDZcQM5rYtp1IiBkNQNw+pwVvx55v2HPHHXTYa6kTUYf1yevzkiTPpIfR3+bmm0SQ+esQstNbsLdZ\nO5o+eQ81NjwJ519Z4DMaNWoUdevWZffu3RhjeOCuG8j+19UMz+xIeMLP9F2UwY2zt5OwYBfLfkom\n97WZHN9+nOQZL/q9h3wtbYXce2z3WK452YV73jiSl5j6HlvYJBGP9Pj4fImtWk1E5EwpbozeXCAL\n+AYYCXTGWe9WRPwppHZegcLEAdbY81fQ+PfND5A4fQgRQ6fSqddQv8e9vcyQtOdG3j5uGNsN/xM3\nPDXtzr8SNn3oJFSBFnnGSew8q07EdRlHbPdY1u46wuhvU+iwfyPJD99B7POTgVPj+zzdwOlR8fnG\nrE0bOpIZi85l0uAORC8bX6Bu32uvvUZsbCw5OTkA9OzZk+//O4Mq3/4VBk4hqktVkqdNptXe/URm\n5pBd9e9svepewrpNpOZ//knUpMkFb8CnFqCnNl7UxaPx7nyPbhxNg5W1OL7SSUz9jYnrPnhMvnIv\nvnzH5BU1rlJEpCwVl//q0JUAACAASURBVOh1dmfbYoyZAwTQnyMivnz/sK/ZeYQnP9oIxvDI5Z0L\n7SbxPW7e+D4kTh9Ct5OrSVw0HQpJ9PAsXOOzgE2+873htmAlxcOJw862Gz8IuMhzbNMBkBTv/MRJ\niP4xNI739n5K3+/+H7Ub7IE/vciLT8/Om82ab0afV8LkHZd3wpW0bRu9evUixZ0sUbtObVatXEXH\njh2dblY3Iaxx4wck2SjOzdzP0chw/t3nKI1bLeP8YXEw7g7/N+CzRJ2nNh7zl9BqeP4VPU53dqtm\nx4pIqBQ3Ri/L88Ramx3kWEQqrpKMbcPpyk3Yk0rC7pQiu2L9TQSIGDqVtVV78Gb16woda3d9/1ya\ndXyD6/vnFh6rp2t0yGP+u5FXvw7TW8Lf+/i9L88YOta+4cS44S16LBvP9Fvb0252HEcPt3C6Tf/f\ndObPHUz/Ny/ikUWzi+1+Tp6/hNStqVx83T20b9+elJQUjIFbOjdm8lOXOkkeFOj+bj7pXn5u04UD\nDz9A4979iO0eW+hEijU7j/DE0StIbT4g73jvCRi+fCdTeJ8nkHGThR0vIhJsxbXodTfGpLnPDRDp\nvjY4JfbqBDU6kYrCp3WoOJOGdiDtRBYYU+Q4rQJLewGdeg3lJne83r5FW/O1+HlW2UjLTONo2EaW\nHHiTsfQvPtaetxS8+FePw8k057H02YL35SZIcfUinBiT4p3Ezz1v1NOzYeYsolpu5uWcQ6QQTnb1\nDxnQ/lKmdTvqJJue2bhLHiOufl1i+0zhleMH+dfP2/IuM+CcKB6Iacx5v+SS89keuNe912oRxDVp\nTGy1CKKBmJEDiBk5gPT4eLrPXE5UlGVy8j8L1gDEHTO5qzHb2v+ZeW7JmtKUKvGMvUw7kUWdyKqq\nPSYi5U5xS6CFn6lARCq0EoxtA6ercsE9/QB3oP6Tf/JbPHfQOTeQuCfl1KQKt67erJPZvNjqFi4b\nekm+/T2JYdeorvlXevAejxZArAkHE4hr3Y7YHSeIrta4yH1j63aDw7/kL4qMmzg9dge8MYr7Dufy\nclQU9/WcxLUd+5A+fQS7vthC1J5HqdGiBgu2baPR5xn0TOhBTpbTChkVUYUPW7bhUOtq/Ksf3EE4\nl056sMC9gpPEpX/xFgemP8vJw1nYTCAjldgXpzkxetUATI+P50+L/0GDjp050GgLCQerlWzpOS+e\nJD0tI1u1x0SkXAq0YLKIFCXAyRX+FFU895NVEST9dCOf5EY4kyrcWnW1gEfP/Qha35ovifMkhle1\nvJux3bxa8nxb8YqJNW5dHMtPHoS6tYit2Zy4zf8i9pdFRH8/2+nq7XkL6f95lOQvttChRwJxtQ85\nSZ573rz1e/ftJDrzKNeGN6B9zynE7fqK9vXbU3dFFTL2VydnRRX2Drqcf9z1MQdPOCNFqhr4R7MW\nNG9dC5MO63pH0bh3J5reEZs31g/g7vDBjHlrLefU2ONM7pjxIhn7nXOEReQQdUEarfwUo06eOYvw\nNSsYnr6VP7U+Sty6uMILVhfDM7bQe+yhiEh5okRPJMSKGqjvu4TatG4T6JSR6rx5/pVO92dGqlOo\nGPgk86H8iaGHVytesUsc7V5J7M4fwYQRW/X/s3fe4VHU+R9/zbYkm55sAqnUBJCYQlWCiBTBhlQL\nNkTxIkX0zgKIBbwDDisquPdDUcFyKCIWbBBAICo1CUFKKCGkkZC2SbaX+f2xmyUJAcETFfm+nocn\nyezMfGdm3eTt+9Oi0YcGu50z6/fozdWQOYec+FTKfqym/Qlf6o5F8fqICAYnXEHmugwyUjLQ5+qp\n2L6Vsh81JHYJR5t+F2s+e5EBW02sGWVgrOyDEpi8I4dvrjw1WfHezp2Zf/vNVH3xFSeVKoLNdm4u\naUNSK0IsSP8JmsJ6oN5dwTv97zjnLwDZSZtBOrR3zvXe6w29bWwsf5+MlAwSPc855oquLFzzCdHT\nBp3fG8bpY6JEFa1AIPizIoSeQPAHc6bcsKZi4tRYIh3LJ7mdOcPSmwgu2UyDLpUAT1HCdFcCwOnO\nUlwft9jbtIC19Tex+XgkcIYw46YFpBbnogfwCyPjigzwDSHDJx5q3I6ePldPRV8Tk5ShbBvWhiz2\nknvkcxocBgxmO09cMY2yJXtof7iGygYZHSu4Zb0ayaogeauLGQodXx88iAsZgPCO4azvezU+u/dR\n9OU3+JtlIhx2JH8/Ym4Y3epzM9mNKADZ388b9u4wbDym7GzK5y+Af7zIpjY9GJr7I1+WOcmJOwa4\nw7zxby6F+yeh3l+D/wfr4CytUVpDjIkSCAQXC0LoCQR/IpqKu7Vr1zCxbBlr105k+g0jgeYCbpF9\nNAOctWxWTODpu+4FoCew/FoJNt3n7Q/nxRO+nR5j53DC7DOHGQfOcLuEJw+CuZrEL5YzYbeOeW3D\nGXnn94zvFU9GRSp6IGpSBiOBE7l6Dh0PpI7NGE52IdVqI7FPEJWqEHQ9XFTuViBZCzjisnDnf1dj\nsLuL+LUqJdMn9eeOsgB8DVZK22ox2U3IDgiwgWw3U/L51xy94trTXMiPB0kkWiD/+hiebZLbWLl4\nCZY97gmN1wUVItUZSMzrztt905vl6p1ry5OW7t2uwhrqzHZS40JEqFYgEPzpEUJPIPgT0dQpek29\nmmDlHoJqViAz8rQ+fLtdCexuu4AFvc2nKljj+jTPx/O4eE2LMIIHzvBWmrZsGgxAXB9W9n2WxTvm\nEqyo5KH3TxBbXMhtJcU8030dl3WY3eyaE0tkZq108lpKIUfqjNy77VNMFRvQWn9E1+NKju4PZl2X\nev5vXSEHDGbAXbZ/Z9cIZsrh2DeVo3CVYwPCgwPQGeBwGxVtnTLVzjA2dRlKSSsO2shgB/oRdjL8\nm3d+0k2ZjLPO3SwgZMxo6tetp92UyehbFLqca5VtS/eusTXOgASdqLAVCAR/eoTQEwh+A1rNe2tN\nRNGkUCEl47Rqz6aTMIIVT7HnvZn8s+5GAlq0UWkqNroeWNK80KLbCHcT5G4jTi/CaCr8WopCTyGF\n6dsP8H3pn+gGKDgUK2FyuoWUUl0HflXecWaNLVVmrWuHcWce99iSKaoPon1+DZVBccQPG0zltyqW\nfLWeN2qqvdfePzyAK+YM4kjtERpWOQkwg1kDmnYxxNxwLTVr/kth79HU33g33+wt45FoK+r33iCm\ny1DGNqkyTr1mDvrGe2mCNi2NDh+t9P4ceuut3hFkrVU2/xItp5O0Nq1EIBAI/qwIoScQ/Aa0mrN1\nht56rfXGa6R5Un8f7ONXEdBKNWczsaFo0S5l/+fuSReNY80sBvc/T2sWSna6f56UeaqPXdpY9J5C\nirBFL5FYKDN2q5PP+ilRocYYJXOwbwDzVluInz4IOSnRPRWj9Ci6dg4oc6K7LpmOXWdQ+eJ8dN3r\n+DFoHNd8PAWrzQpAhFLJ6q4dWf+3LnQfcif1xzPx0QZgenctywcqsF5Wgl76ltArjjC1026Ivpbx\n+f/m+PsqjLvyeNBPja76siataM690rnyxfkYd+aBxUD8ex+d0zGtvyen/ywQCAR/Zn5pMoZAIDgH\nWh1S32JyQyMZUQNIR+sdHXY2GkVFo0vYOIkBYFZyPeoPxvLdzye42/YEuzyFGHQbAX5h0G0Eu1wJ\n7KkESnZieu9pjn9SRa5BSwYV5LzUCQ6sBYsB/Y4XyCrNQp+rd1evtlez/0oNE7dG0rHMTpnGTmzu\nSTofrCf0g3XumbWDXiM19iq02lLirypHW/AG+TESj16VT8zzmfQbeT9WmxWlJLEkOoZNCZ0JdyhJ\n/Oowmccz0Q/VE/KzAa0F7t7k4sH97jFndBpMTtpYMjZMI6d4C7rudd6JFY2taCoXLzmv90fXvQ7/\nthZ03et+eWeBQCD4CyEcPYHgN6A1l6fl5IZGvKPDFKsgqflM1V+iqXP4aMV8kq07qf/pBTZbngA8\nbmITR29Rbjei61No75NP7W4FxmNWSqVwslIBPwn9j6+Dy0GGjwbCdWREDSA/8jL0/+jH4PjB/Nh2\nNaHfKViZdgITEKjyYXCnInI/Wcjxdz8ls9conjEVUP9TLbr+YYy6dRTHNh3zXu/I2Cgmx0cSZZIJ\nTwngpCWU/EGB3qII3ZTJmPfuRVtbi//BGPj7fIjrg/6L28jCBG3j0A+ZS7wn9N20gOIX28Q0QXvn\nXOJjTw/zCgQCwV8dIfQEggvEv396jb012zGY7Xw44q1TL5znFI2muX5NQ7aaipnsWT8fY9p0BpQ0\ncRObnH+6KwF1RS5B1jqcsWbMRSHEjBlLuv8mMqr2wZVTySnciN5ZQUZVJanZq3jW7MOAT3P4clAO\nuyONFPT1Z9xXTpQKJXGDg9Baf6T0zWI6Fzips3xENlFkHyznic27cbq7pRAWpuHlHvGEuXyILXGH\nbov21KF+YTrPDrrFO29XO3AGcW8scffBmzIZ4tz5cw/mlXPz9zbk5AB2DUqgp+dRNC2gWPTWtnNv\ncfI/NLS+WDgf4SsQCC4dhNATCLgwfyRtJwfjsNRgcwxu/kJcHw50nYztvZlohsyka+8hZz9Rk1y/\nnnetdouaou1wYAncOZ/kuD5cW7Sdhs9Gcthg4ef+k/ja4yT2jAyFO+fDpgWUrLWhrD2BYvN29Mu/\n8J5ev66ciu3VlG1tQ+LDY7n+nQ8JLID2GzUYfBREOLX4lxkAJ/47ZD4e2ZPPfA2MVDhZ2c2P7/9v\nC3X1DQBoJYn34tsRc2USyybGELfhAG3WlqOUwd/k4Nhrr5Ay6BZOLpuFeeMRygoe5e0rLyNj3hTi\nmxSmhO6LQFN4gnxzAIu6nipEafo+iaKI5ojefgKBoDWE0BMIuDB/JGcNuY5F6zu1KkRs691h17qv\n7oW2nzTvd9eS1hzAloUemxYQUJlDZ+CZA0vY42uH0mz0g17zulmrzLeTaID8/go6Zmdz9OUFfBVf\nw3W5VfiaVMSXWCl8cxXfdTIx5ABYG+pof9SKb3Ic6I5jxMWyy0v4TA7CEWVjY84RSr9yt0tRKBQ8\ne/vtjNl3gEq/YEpvnILpeCjD8ubgb4Oy8BBKwiF+2sMA5GVpiDrhy8kNBgb8+D1rRhlIfeBD7+3p\n/jGTuhcW8VOXoc17B7Z4n4SgOYUQvgKBoDUkWZb/6Gv4U9CrVy95586df/RlCP4gfu+w14Ed64n+\n6l6C5Dp3wUZjWLFoO4ZvnmORfTQ33DDyzNey8x3InAODn+GAHIvquycJ19jIQslb8W3xs5zg8dJC\nUmOvgrtWk1ORw8IdCzHajfir/Xn4vXoCsw9j8INgMxS3leiCGUNaLPsOmEguMGHUQF24gvCp95DS\nLpKMnf8iy88Xn3fL2b3xJI2/OfoG+PPMkGSipz3MiwXt2HyoklCtmhqTnaU/LSH2xFF8k5O9LU9y\nN3xExfznMZkDcMkyXavKqU/rzLKJMa22nGmKCE8KBAKBG0mSdsmy3OsX9xNCz40QeoLfjcacu24j\nTrVAievDrsIa1B+MJdm6k03OZJa1f4Hl9/X19t27ps0dfLSjCL/g1Tx6PJtDLjOvRkRwW7WSdEsR\nL4VG44jrTl5lHmPq2jPx6zJq7ruFN4KOUmerw5Kzh4kbQGVxopQUGNUuNqYoSD+kIHjMOKLKv6eg\n3+2sz/6Ue5YV4G+TKQyH+iDIuvJeYtsYmDVtPna7+3dGqErJ8oQO+ISpia2UKYvTEhXSDod/Kfba\nNnzc4xZu6R1H+CcrqBpzFy+X+vBItBXX7ClojQ4OJ4aQ2WsUg3d+yv7kQGJ3FZE/KpVnmzh7vydC\nRAoEgouJcxV6InQrEPxONAqJ11z/JLhks3tjkwKBRevzaai7kdkBLjZH3s30IYnsKqxh2oYF1Ct+\nZk9xLTUWGyrVIRb6K9nvE4YDFx+GuMizBJPt50JblU8n1PT8rhTjURPH3/2UrDENqOztmLHFj04l\nDZ7VXGR3gKyeWiaOeAzlO5+wqn8k+Q0byAs9zi1hfvifMBFbDfZyB49seooSiw1w/9J4MSqKlPZB\nLLpBRY+TEmGb7LQx26BoP2qVE5XDwD3HXyT+prfQvrmU2Z7CiVs+fodORgcmfxXxDz+CxbmBWe0a\nmLfaSecCSN7qggdaf26NIcnzEWPnI97ON3wvhKFAILgYEEJPILjQeBy8tfU3sfl4JIviR/N0J7Xb\n0Wsyumz6kEQWAfKQCTzdLhSKtrPnvZn4SgmY2vozput1rMuWsNo/IvRkEY9l2fi0v4I7LEqif9BS\nebWSg+2slDqdvH2FgrvRsr9dOrPfz0RxWTlb4wcTYt2JVq7BbLcRXufgn2+YcRrnoLG6SDwJ+VPT\nGN7QEf/awwA8U1rK6rpTvedGBQUxJyYalQuohic/Ad+YNigsJfi2b4eyvRb74VxstSCbzFQuXkL8\nm0u9Iq1T8rX4v1dAu+l/RzvoFjIq3Nvjpw/C/4N1rc6dbSrAgPMSY+cj3prmuJ2LiBPFDwKB4GJA\nhG49iNDtpc3ZxpKdF62NPVsxGo5kYogZwDTF7FPiwbO9WY6eh6Zh3DuiOrHH1046WqKNGSw7HsmC\n3CdJKbCypwPE2iGsGI62C+fDSW3JN+ZjRiZAHcDSldEoD+7DqIE91wWwLj2ZvTXbeW6VL10ONXjX\nM/jBG+ODeeQePUX33cfRXWU8WlZG4xTZDmo17yV0oCJSwcYUBUNzXXQ2+CGbzDi7daKIaqKnPUzK\noFvY/8FSXEteQxEUTbt/zW8+cuws93wmfi9Hryl3exzIAQm6M4o44egJBII/EhG6FQjOA/22BWRV\n/wyWWvQ3/ffXn2jTAnKKt6DfUELGdf9xi0bPGLJgawXLfWaBYj7Q56z99NauXUNXVxUvtG3PyJTb\nCDz4CRmlR+mgW83hhNnsd4bRqaSMkDr4rJeCVLWL/P5m/q/Tk+xd+DRbtMVcu8eKFHASAH8bRO4w\n8ciRQhangerWqyl+93ucpgbsPgo2j+nMI+Pm0NbVln4bczEbnQD4ShKv3H0j16nasFx3lI65pZRE\nKDDrnyXe1JnKxUtY1qeBT3wKSXduQM8tzDb7s+/2zlzmO5ZPPCIvd8NHlL72CtF3jiSlU5N7PsM8\n4Ka0bEZ9Pu7Zrx1Xdi4VrGIUmkAguBgQI9AEFw+eJrsUbf/NT51RYyDdZCajxtD6uksHu//90toD\nZ6CP7kgWJvS5njm2cX3ANxgqD7jnzG5a4N3dYLEz94uf2VVYgyk7m+P3T8KUnc109Wo2hDWQ7efi\n62+/ZfqXMSTSk+DhT7H8vr70yNMSYIP4KrjisII3btPQffCtVM57ksDswwzPsqCqN2OtOkl+FByM\nAll2Ebi3hAdXOdhZuJGI95ey5tmrifvv+8zp9yxTU68nKirKK/LuSGjD7sQuxFefQDdlMsO/qSGt\nAO7a0o5xXcZ5mxf3atOLpz+SuNHkHsGmichEFXAITUSm9z5LX3uF9vtrKH1vjdvJaxR1jW1imjyT\nPwMtR88JBALBxYoQeoKLhwsoClKvmYM+IJnUa+a0vm7JzuYi7UyiM64PGdf9h/TodO+YL4q2g8UA\nuq4Q0+uUm7VpAcElmxlQtoxF6/ObzXENHv4U6SYXIU4nd20uh7z9VB6J8wqkxGt641DLlITDt30l\nanGRWX8UXezPWDQyCsAhwa5kGa0sE2qEmEowq9ztVMb+FIr6rkf522NbWXHndAJ69CCrrAyAXn5+\nFF7Xk0l92+JQgDbUSeXiJfhbzBh9/dh6g5bcDR95RanynU9IOmJH+c4nADxxxTTSo9N54opp3scS\nPe1hjnULJdrTR8/LGeYBX2gaZwbvKqz5XdcVCASC35sLFrqVJCkOWA60AWTg/2RZXiRJUhiwEmgP\nHANukWW5RpIkCVgEXA+YgAmyLO/2nOseYLbn1P+UZfldz/aewDuAH/AVMF2WZflMa1yoexX8Tpzv\n6LDzwFSppvL7cHRJarRxraxrMTRfu2nD4oEzmoUfUyNT0Q/1uHlF28n5+Hb0fhIZNhkGTkd/YJl7\n/m23EdiLd3NUew3ThySi63FqjmuOj8QbkRHUu0ysGBDJ3/PjmhUq2DNXobZLVARBjdOPWf81ciR1\nP0d/0uJjc+fduhSQViThX37qVowaONxBIslYRc7RWu48XojJk6cbqlTy+YCriaEaXcdjVGRFoHKB\nf54F3VvutVf2aeAbn70Meu0VNPtrqKsqQ2W0Uhyv9Yq4ZvfvIWXQLaQMuuX0B/8HjSYThRQCgeBS\n4ULm6DmAf8iyvFuSpEBglyRJ64AJQKYsywskSZoBzACeAK4DEjz/+gJvAH09ou0ZoBduwbhLkqTP\nPcLtDWASsA230BsOfO05Z2trCC5mLqAoaHTTAE4+tbB5kn1cH5h0KgxJ0XaoKwFNoLtytuWUiqZs\nWoDeTyJL6weYcWyZyzatGoCMskL0ISoy/HeQ2i6UHL9CFo4yQunzUAr1LhMhPiGM6n4XFVkfUb36\nIRbsUDPBYCP3CgfdlfB1PwV3bK0nsVCmw4kKlGYACQCNEyoU4COBSgYXsGKQgrLuCnbNyONotTtE\nqwSei4thRKeOKEpLMSXEoO3bgS99yxieWcs3Q8LolZaGbspkbnzqSXoZNahuugH/8GMcLsol9riJ\nvA6BhHUaev4P/hxy9M7E/1IMIaZICASCS4ULFrqVZbms0ZGTZbke2A/EADcD73p2excY6fn+ZmC5\n7OYnIESSpChgGLBOluVqj7hbBwz3vBYky/JPsrt0eHmLc7W2hkDQKropk/Hv3x/duGtQfzCWhsM/\nsGh9fus7b1oAJw+Ard7d8LjbCPAJAkMJB3asbxYSzEkbS51vIJc7JO6tNTO1too0s4KMlAwWBqrJ\n0vox3eFgV2ENC3csJK8yj7zKPA7XHEXr6sBNbWdjXPwZUccLcGyu4IEPy9B9eZJx9Qbyezt4bKOV\njgENGDVgUkO1FpyATQGF4VCm80UCrEr4eISOTbuMfPzAXq/IG9I2lLwuXRmpDcRWUw1Aua0O7lrN\nl90DmfSwik8SfDiwYz2VT07C/3ABncpM8PkPxL+5lPAnHiWvQyDvdB7mfl6ekLbp2w+8od2z8j+E\n4xtduTO+T2ehZQ6eCOUKBIK/Kr9L1a0kSe2BNNzOWxtZlss8L53AHdoFtwgsanJYsWfb2bYXt7Kd\ns6whELRKY2EBK0aTbN3Js0FgHzKh9Z1bhnI3LQBrHVjrsK2fz2bD3wF3SFBftpk8hYN0i5meLgmV\n3caDLn/0uXqMktt5q7Cq3GKl7aklzE4j/QqPkvbOwyhNgZSERxLirCKq3AkokNcHcrMkobZJyGUq\n/GV3Za1V7XboXErIHKTk7lUWlDJsrq3nuRcP4HK5zx+vUrGmXXt8gvzRRMWi8PfHfm0fjq39xBuC\nHbXPzOD1DlZdWc/Phrnk9TIwrliNwiYTFewLuEOyjk5D0TW2P9l0HxzJpPLD4xiPGt1rvbn0zO7b\n/xCO/y1dORHKFQgEf1UuuNCTJCkA+AR4WJblOsnzxw3Ak093QRv5nW0NSZIewNOHPz4+/kJehuBi\nwSM4kgfOgLgzhAM9odxdhTUs+i6fWcmT6eoRfprUmQzYE+htunvy+ABSgivJkIzs7X4V+oLPqAvS\nkVeaRcfAbqjs7UDhIqlTLdde9jjTNkwj4kgNd693El8JfnaAerg8lrrLZYo2nCS8WkZrU+ALyMhI\nsvszZVfI+Ch8ABtqO4z/zMG+djB1/WEqnW4Hz0eSeDMujp5+WlzISCYrdtMxePYVknsPIfn+f3hv\nc9T6KiQz3JNVz78HxZFrrIFJ7ZmY27ZZvmCzNiOe56e7cix8vJHd14zh5rnfofPXcOikkTqznSA/\nNcOTovhmb5lb+DUJd59POPa3bG8iQrkCgeCvygWtupUkSY1b5L0vy3Ljb/NyT9gVz9cKz/YSoGka\nfKxn29m2x7ay/WxrNEOW5f+TZbmXLMu9IiIift1NCv5aNOYBtsgXW7lnC/3evo2Ve7Z4tzW6QPP2\nBLpz+CZl0rX3EG9IcNH6fHYdCkZheIzUu79Gb9hDFiaQIT06HVXNSCxWDfgWsabgbVIjU3k9chqz\nVsl0KWsUeWD1UdIu6SRvR5Tyj3tV/OtWFQ0a92sSp/7HSemSwGrDo/t46Egxt317kEqnEwnICA0j\nO7ELcbFaDH7wQy8J/7YWrKlVzN31ODkVOQDeNi/hN12Pwk9B7IMTeHTAbNKj0xl563PEv7mU/WHt\nWw115h46xterT3BIrSL+zaXMOaKgxmTnRJ2VAQk6kCQ2H6rk+W8PtBp2/V/Csf8Lop2KQCD4q3Ih\nq24l4C1gvyzLLzV56XPgHmCB5+tnTbZPlSTpv7iLMQyyLJdJkvQtME+SpMbfwNcCM2VZrpYkqU6S\npCtwh4TvBl77hTUEgl/Fol2LqVf8zKJdi7k1+Srgl12gWcn1PFrxEprkmeRU5BC45yAvrbMTrSik\nw4tz2J/Ynv98foTgI0e4da8BU2o2octWYTS7kFROioKVOAEJF5q1VUyTJEb6OXh7iMT8WxU89YEL\nX6e7QkkCJGSMGol/+lTz+c4KGm3swM5+fKuOIcSlAglC/n4vb2h2k1FeirOkmn9HBJLnZ0efq2dS\n4nzKZs6j07G90L8/XbJ/BiAUvJW0uwprmLBsO/VWB3UWB2umpHvvubFf3rHXXiFl0C08Nqwrz397\ngMeGdWV833ivY9fU0WvK7+2siekWAoHgr84FG4EmSVJ/YAuQh7vgD2AWblH2ERAPFOJufVLtEYav\n466cNQH3yrK803OuiZ5jAf4ly/Lbnu29ONVe5WtgmidUG97aGme7XjECTXA2Vu7ZwqJdi5nec4pX\n6LXEO0YtagCp2avAUAKVBzAZQvj+UDhWo5kunsxROdCPt0fKlETDrBV1KEs1HGmfxJU9SzBmlaFL\nc/LW8EnoXv4vqcdczdYxqUDtdFfSNvp5MnDYYmH88UKMns90sELBy+kdSLvOwJYSHQN/8mHTjfFc\nNXkuHT57iuCSkrXIcAAAIABJREFUzXzol8TcoCD8fdS8eeNcXvrCSsW2nUw9volB82c1H1/moXE8\nGEBqbDBrpvb3vuadgOEZh/Zn51xGnQkEAsGfkXMdgSZm3XoQQk9wzpyhJUjGugyySrMYU6xh4rpq\nAi53EBZeR8F34ViqfagIkzFpJNpU++Bns3Iwyi3UtDawqtWsjevPQ0e/R2O3otBItJkygX27llG3\nQ4W/VYnFpSK83o6yxUfW5nBwa1ERB21WwF2MMbdNW4ZEheAz+15SDr8B5moy2nUiS2EnPaw7SduT\n6L/5I14dFMGO9hUkhfbhwxFv/fJc2aLtGL55jnnGmzmo6spTN3X/n52wP9JVE46eQCC4WBGzbgWC\nX8kv/vFvbAliMbhHm3kEX+MkjBtX/ozxhIlCkx+xWh9cdrcyq/eRmDlBxfC6RO54L4/OJ2SUnoS6\n7A52hpz8Capc2HD32Stf/C4n/VUsvVZNpwqJu9a5RV5jqBZgXnk579WeypMbFhTIjK4x+Jth47AI\nBveaxNzyCKaHrSbDaQBLKRmSAdXRwyjL7fxtXyiafgnea29a4DBz7vsMXf8xq0rG0fPpO7z3Hlyy\nmX93UsNd95//s2uFP7LiVcyrFQgEf3WE0BMIGvE4dWvrb2Lz8UigFeHhGWfWoEvlREUDne1uFzhn\n2NPesG21eQv+gJ/FiqVOjaSBsrYyy4aoCPEJ4Y5rRuKzNA9kCackc7itxJfpSp5w+VBcb0ZX6xZy\nrlol7SthxionvnZQu9w5EApgc30900pL8NRrEKNS8VZaezKv01K63UZaAQzLqWPu+q/ZJ39BruJ2\nPrm2M3qPE2m6XE3l4iW0mzIZvSc86w09p2SQGpnKzXlfE1xxkIS8rwGP0GvRDqWlsGtNtP2S+BMV\nrwKBQHDhEEJPcOnSMgTrceqmx9g5nDC7deHhmXt71KcXHxjTmeVTSlC3Eehz9WSVZkFpNg+mVnPM\nN5yYa4ejWP41LrOLSIUdP52WqZfNIzX7Zfb0qMe5OwCXVuan/koeV5aTEuLP9iANvlU273IyEGh2\nt04BiRqHg7HHjlHudACgkSTeiInhxKBAnrpaydK8CipMIdSFK9nReyyaiExUNYfQhGZiqux2asxb\nY9/AJui3LSCr+mcaDJUoDI8RkjiU3uUN7Lv8Oq5o3KnFdJKWwq410fZLjt2ZXDURVhUIBIL/HSH0\nBJcuLUeXeVyq4IEzWN6ivUqj6JiVPJkgg4WFJ4cz3W81Qc46cvatpC48lI7BHamTZeS0MK77+xzY\ntABTejm7D7ThP1f7EWdysChrHoaYG/i5fT635zoJrpW463MHHa6SISWApHl6fvz7nThsLiIM4OfW\nc8hIZJQUsbnB6L2miWFhPBgTib8NHNtlko47KJNDaFcOpXFB3BW/jf4db0dfpiYjJYPKWYu9Y96q\n501p5t4BZJSXgt3MjRXHmFpVSWpcZ2rvmcnwpCjufmtbq4KrpbBrTbT9WsfuQod0hZAUCASXAkLo\nCS5dWk5lOMss3VOiQ0de9cPUWO0kSr3p7VeIXqsirzIPjcsXm8LCNJU/r/loSO02AsfRHbwaqWLk\nDzbW9JepDyvmnZK1PL5JSbDZiUMClVVB0RYdcemD0O59Dt3d15L/0bf4OGT8DLCypprnKiq8petJ\nPr4sj42loo2K7zpL3LhDRu2ChDI4Hg7ZHWB/72oGl/xMaraah7tOxvbWPzAOug1/3OPe5jU6kJxq\nm5Iqq9GXn6RBF8OABJ1XADWtsm0quM5VKP3aPLgLHdIV0zAEAsGlgBB6gkuXVoRdTkUO+m0LyKgx\nkHrNHG9V7Q29bRxWrOCGnlO4M8aC/08v0CXYRY6lgZr6UhIKI5m2pZx3BjrZHW9E//XfyLD7oI/0\n4f7VDuKKIMJh5+nbVVxeUUx8tbtAw64CXKCyShTOew/fQCuy7SBJDTJHLBZGFB2n3jO3LECp4P0O\n8SQofXFKEF8FIFMWCtFV7g+zrIEVo1VotZHkODvg23UyVWvnkk4uewoh/k23g5lR4S6+aCzCAGD4\nfNi0gIAWjuaZBNeFFkq/RiCej0sncgMFAsGlwAWdjCEQXBQUbYcVo6FouzvXrvpn9MZD7tCuh43l\n71Ov+JmN5e8TWfEGK9qUshITU6Oi2KewMeGHMtqWOPl7lg99HSoySo8yDwNZWj/WDInEGa4Cu4Ko\nMgVX/+TAz+ZExj39Qun0LGIHS7UPcq2NsccKuKnwGPUuFwrgmchI/jM0kS0TQrCqZW+Llegqt+BT\nASYfDQNe+oB27dPJd1Sgj2rHvD2BvGgdRRYpaIbM9N5PamQq+qF6Ektkjt8/CVN29hmngpxpasT0\nIYle568puwprWp2a8XtwPpM1xDQMgUBwKSCEnkDQmKu3aQEZKRmkh3Unwz/hVEgXuKbNHQS6utNR\nPYpXjhkZ8LmaTJMNgySjdUm0616Drb0Pn47uhsM5kHnBsRTK7jllP+usqDQysSckHl/lZGcXCYfk\nrqxt2ioF4PmKcnocOcw+q7sn3mD/AHZ368qVncP4dlAwd9z8GMW6U0eoAIfkDtkuH+VEm5ZGRkoG\nY6xJTF1RwyPRVgI696N49L94pXaVd8xZI5Uvzse4dSuVL84/r0d2Kmexnp6b73OLZQ9/1BgzOLP4\nFAgEgksVIfQEFx9NHLjfhIEzoNNgGDiD1MhUMjqMQC9Xk2M47N1l7Q4NlxdpObh/OtdnVZNWAI+v\ncpJQLKN2RiKnpfL8QCWJn+Zgqfqe/f4OzGoLAFHOGqypVZg1EGyGe9bjnUfb2Pt4q7GB1PwDvF3j\ndsGiVCqyOnXmtdhYNC5oUyfxxCon7X5WUjEkmENRUBoGJqUKlQxmDeyLk8hY5w7Fjt1Qj3L7HtTv\nLmT5fX3ZWP4+WaVZ6HP13nvaVViDIaIM/7YWdN3rmm3/JUeuUczZ1s/3iuRG/kixJVw6gUAgaI4Q\neoKLjyYO3HlxJoHYImS55rMXGfDfOj5Z/W/2zB/MgR3reSTays1ZP1BTreKnK5wYNRLBZpjznpOU\nrLY8rY3k+u9NpBXAyC3u9iiyBH5IHFL78nqwDoUnRKtxuIsnAEwOB4OPHOaB4mJsMmiAZTGxrO/U\nmYpYFU7ApgSlpEAyNFD0wvNsT27PkxNUPPw3FUU6d5qtBNRJeMXcqsvLyO4Aqy53z1zLSMkgPTq9\nWU7eovX5POM7mtphHdHeObfZ9tYcuaYCsFHMaYbM9IrkRn5JbLUUkn9kqFcgEAj+6ohiDMHFR8tq\n2XOlZTuVMzA2LwplwRESbQ6Sr9rJnvXzCT/SEd9Cifs2O1h3c29Kw3eTUCajkiHjp108Ex/N7sRQ\nOpbVUhl+GfNWH2PztVEojhYxaoMZhwQ+HqEnI2NVSjxeVMz6+gbvuj1Tw3jX2gaFy+3ztTXA03cp\nGbvVSXGEiwF5sKZfCNe0m0JW6QIcLhfrUuKJzPuBrKtkHq2uIjMizi3mogagj11ERs/pAPgWVjJ5\n5yE0IZXg7gXN9CGJLALsQyZAXKg3HDs8Kcr7elMaBWBeiYE37+l9qlCi95DzehtaFnGI6leBQCC4\ncAihJ7j4OEsblLPSikBsrUrT+cDdFL32CoqbhrOn/Hs0Q2aiGx5OndnO1i5DKY/dwDtDJGZ/6O5z\np5Jh7PZSVJKCYIvM4F17wSXja4gg6qQVpRNvaxSA1bUGnik/4d3WTePDsoQ4jrZVoSiQsWtU2HDw\n4dUKxm51klYAPjJMnebHeIOR3RsOUn9iMqFaNU8mvkJ02HGSdakEBCcybuAMiEyFyFT0SeO9a5q+\n+xc97LvZ+uVcdkX2bNVtO5vg2lVYQ53ZTqCPihqTnUXr83+1KGtZ7SqqXwUCgeDCIYSe4NKhFYHY\nmrh5w7mBrJH1JIUWopbnMYt6tDsfJWkgzB+eQI5Pd/R+apY7OzIw80NQWFjVX0mvfBfdj+J25JQu\ndFWFKJ0uHBKYfCUMNWZuKTqOobFdiiTx3/h2tPf1paCnL3EFLmStRJtkf05aQqkIg/e7tSPCvAFd\nnYvZK8zkXFXPrT7vU5awgOFJUXzzUy/u8tlHwBX3Qq8JQOvidaX/HdRVOXjVMYoAj0g7l6kWTZ9T\nTrGB1NhgkCTqzHZ2Fdb8qly4pm1Tfq+mxaI5skAguFQRQk9wSdMoappOf2jMYzt5fAA/Hqrk0YqX\nwOqeaWta9g/Cirry0pTJ7L+nPfNiYvGrXMTd39TR+UTTpFcZX7sTqxJkq4MHCorZY3UXZyiAWRER\njA8LRwaqteA6bkFX4T6yKtuM0lrNuBoNS/tdThuFA3W1ikgg/EdftPOeZHpkIve/u4OX7RtRK2vY\ns24F9oibW503u6uwhoOqruwO/Sf+GuUZnbSWfeuaiqOm+zbN4Ws8/68VUb9X2FaEhwUCwaWKEHqC\nS5pGcTN3ydtMLFvG2rUTeXryvei7TsRw7DkWxY9G03smrL8XrHUc/dGA8thWckt+wPLMPUTEH+X+\nzwLxLztVteoE1GpwOeGNsgqW1lR7q2v7B/jzemwcGs8GCQg3QWUw5LtT4zh4mczlh128N8BJmO5b\nVveHtC3uCt1Ph/hwvY8PC97dQY3Jztt+txCkUvHPuhtJXbuGnoFfMCt5MnCq6vW5L372unFrpvY/\n7d7PREtx1LhvS4H4v4io3zJsezbBKcLDAoHgUkUIPcElSzPHSr2aYOUe0tSrgXth0wKCSzbzdCc1\n9F6NqfrflP/7earsDszh4LK4+DJzBbtj4ZZ6F/6capWiBHbUO7i/4DBW2b21jVLJx+07YGyjQll1\n6hocQG2QjITEfwdL7I1TEmIOo7+mlHuztMTfdxuy30YWdjhBgWylQWll/47XqTHdTaCPkocm3InM\nnQSsz2e6659wZDNdgeX3NQlRS1Lzr+fImcRRS4F4riKqNSHWmtj8tQ5ha4Kz6bmEkycQCC5FhNAT\nXLI0CoPO1n2M0BjQt+tERqcrSV0xGrqNcO/kKdyo/HgjllILkYBdAWoXjN7qxH6LhvDUSuTyECQX\nGB0ORhUeo9jhAEANvBYdw1WBgUiArgrKgiCiHk4Gwg8DAhm1rg5dHTyy0cbLt6o56hOOsjSKzkd/\nxnflLozqNvTrmM6Udt+zIiKcY0XXUwF0igjwCqHl9/WFoqfclcUtqpGfuvEyr9g5H851BNm57neu\nzt+vdQhbE5wiZCsQCC51hNAT/OU5k0M0fUgina37mGmYw7QQFVlqPyj4jJd2HKHyw+Po/rUUbVwa\nK/dsIdu/gDu0vjjtVnzsMnV+8G5/BS7ZRUJkCmVxDh76cTtf1dd7z397SAhPtWnrnX4hA1YlSCqZ\nt4cpGJwrM/K7ehQ2CTTQNdlFcFhnTMYCvh6cxMTA/lRUFOC/p4SQsmN8rL6em9euwXGHi499WmlI\nfIZq5HMRYr9HscK5On+/Nsza2n2KkK1AILjUkWRZ/uW9LgF69eol79y584++DMEF4O63trH5UCUD\nEnSnC54Vo+FIJrnmCEr3hxFzzQB8ln+Ny+zCv39/4t9cSr+3b2PqR7mkFYAxMYYyuYYNncwMzJVR\nAofKjTy2v4jGkbUJGg2r4uJRq9z/H9VyzBlAgwYC3H2VcWokNt08nqnPzSanIoeFOxYC8Hjvx3n5\n7fsZtsnIZ+l+3Lndh875tRzp7E98ejCaITPpep497H7VMxIIBALBnw5JknbJstzrF/cTQs+NEHp/\nfj7Ydpz5X+0jRKsh3F/DUzd1P8192lVYw3Nf7gNZpr3Ony9ySwkL0BDip8GmOorZ/xvGdprID3v9\nqXEdwuXzMfdV1hD6bQCJFcUY1FqC7SYMGi1zrpjI0YgOJHWsxVXyIvdsNCGhoDgarsp2UWWxMbrw\nGLWedilahcSKuHZ08/XFpgCNq7W7cFOthUC7jMJX4tG0qRwMa48kgUsGv7i3UAUcwtGQwF2mEr4O\nr+f66kC2FF7L3Ye+QdkVxoXks8mZzCtt5qMo2cFDqtUskcewX9mNtPgQtnjClTenRlNttHFZVBDL\nfyzEKbsI8lNjtDgID/Ah3F/DLb3j+WjHcYw2J/4aJVd0DGflziKuTozg+/yTPDasK+P7xp92D+fj\nAp5p3w+2Hef5bw+ctoZohyIQCARnRwi980QIvT8/aXO/o8ZkR+FXiI9uPZf5juWT+8Y326fRmWpJ\nqFaNJVyPKuAQmLtQf+zeZoKqXe5g7sj/nO3dbPTZr+H9xBHkx0j46NbT4UB3bs/bTpBUTGKZC6vD\nwX0lxey2uNulSMBjuggmhId71ysLk4mqdr/aNHQr426vYlaDnx2cbRTceKXbwWu8L3t9EurAvVgr\nhzBU2ok9Igv1yXS+NY0BoIeUz0Oq1bzqGM1uOZF31AsYqNzDJmcyE+ytTwtRKSQcrtY/66FaNTUm\nu/dry31DtWqyn772tOPOxwU8076N72nLNYTDKBAIBGfnXIWeyNETXDQ8Nqwr87/ahypqIw6fQ2hC\nM4HmQm/6kETqLA6vo3doww9MOJKJeuID6OtGUe34krFdJ/KDw58a503U2j5BUlrIj5GYl6ZEpa1k\nfXIcqhOd8NH9H6qAQ0w8Ukz3E/UcCGvDbHs+nx4t81bY9grWsjQ2DotaidnhwqQGHwf80ENm1A4X\nCoPKG7aVOBXCrQiChkD4MF2FX+RL2E+MQa1b7xaigLnoPhRATfxS9ms1XBaRDYVjSIjwJ6+6i1fQ\npcYG82rJaACWyGMI9FH9Kkfvm71l7gbMe8u4LCroNEevNc4n/+1M+z42rKvX0fu15xYIBALBmRGO\nngfh6F085FTkoM/Vk5GSQWpk6llfD5u1GOPWrd58u5aMXzuevMo8Lg/sQMz+ItK2Wsi+MZHnp3xG\n7oaPKH3tFWJuGMOxz77n9vXrMHlcPJ1Syer2HdB58vDq/SQCzbLXqTOrISsVrtnhbrfixN0rz6oC\niwbeHaLkUIzkVX7pYd0ZHNGDVw++z0Nd7mDcFY+772XvB+h3uWfWpiaNP+36BQKBQHBpIkK354kQ\nen8dnv2/20n8NIf8Uak83vtxKhcvQTdlMtq0tGb7mbKzyXz2b7zdx4RvrC+z3q1BWapBDvTD8bc7\nsC95G0WdlTEVZRTUGwFQSfBy22gGBQciyW6V5gCqeyrx2+dEbQNfT1WGA7dl7gRKwuE/1ys5Hi1j\nVShQyjJOSUIpy7Sz2Znj2xG9XE0WJtLRor9n2+/2vM5I0fZT7Vri+vzRVyMQCASCJojQreCSZexW\nF8oCSN7qIn+khP5WJRkxEi29v8rFS+h8sJ5JilCibp5Kxz2zOVblQqo343h9Gf8sKGZN3amJF5Mm\nTSL68jLCvjqModJFSL1b6Pn6O+lQYsFo9uVkMKgMbqOu8cOlBOKr4JatLpTDLGT6qhjc0EBmQCAZ\nl91DauFOGDiDDMNh8Lh3fwo2LYAjme7vW2nbIhAIBII/P0LoCf5ydHxkhtfF+3uunqzSLAD0Q/WA\n28mrXLyEwKFDwGIgOrGIN3a8QMbN97Mm7BOqPi7jjdwyb7uUmHYRDH9tBJP7TnafJ+opUhblg6dl\nnjUoFHM3C/ZaKwF+AagMbvfPrAZZhnp/qNXCZ+nw1uE6+u4NRJdkY1xKLAyY5b3u1Lg+6P9M4dnG\nxssDWy/wEAgEAsGfHyH0BH85tGlp3ny8jIoM99eUDK/Ac9bVYdmzBwBd9zpyN9VQMUDFswfeZdMr\nRzDY3BLPV6ngigfa4UiPZFv1NlTbGnjQfzQ3vHWUtckyQzeDrx1OOs1YygJob7HhLDdSGwQNwT68\nMdDhzsNDBkmio9VG5fpAjCd8QVISn4I7PNokLHrW/MPfO5R6hgbMAoFAILh4UPzRFyAQ/OYUbXc3\nQi7aTmpkKvqhelIjU6lcvATj1q2Y7EaOdQulZvxQCjKNdC6UKPrXUT5bmI/B5kQC5s+fz485K+jS\nN4InjrhY9LaNaS/vp+rfL9C5wEXsUSXHde7ljGoX0dMexqkApQySHd6+U0VBrMJTaiuBLHOnVYEu\n1Y5vGwVOWYspN9ct3Jqg37aArNIs9Gsnuu+jKY2h1BbHCAQCgUBwJoSjJ/jrcYbcMt0Ud+h1WZ8G\nPvEpJN25AZ+fD/LaiZPedikpof4sXziHkJ8OEPjWt8zKrsBJIJYTCsBGm3YRHO4C2dcE09ZYi3mr\nkZJb+rHz8KeMDVNhN9j56GoFveKvYVTeGp4P0GBRSMiSxKcaJ+NC6jH6R6A8auGoNo6kJ5qHRTNq\nDGAyk1F7wn0fTR01EUoVCAQCwXkiHD3BX4+BM6DT4NMEUWNId+TIJ0g0JrJ85HJe9Yi8MKWS7zt1\nZmXi5QR+tx3j1q1UZDW4w6wh8cj+fu6TOK2k2F3MsnUncauRj/orWKXOJu3DHJSVDo7rJNalKfm0\ndAvj+j3J9uITJFltpy7CL4xV18WS3QFWjfT0jvO4jwCp18xBr2pHqi75dEHXGEr9LcK2TVxPgUAg\nEPx1EY6e4K/HWXLLbDYb4weOZ//+/QCoVEo++FsPunQagGLpWly1tRAfj29yMi6jEUX7joSMGU3B\nshex+0JkRSVGG5iLv6KzTclYnHwU6yRQEwjU0xYVIU4nD6GFXhMwVWmY9fICVvWTGRndFsa/wEgf\nDQs7LgRrAzkf305qnWeSR6OIm5R54Z+RqKgVCASCSwLh6AkuOkzZ2Ry/fxKm7OzzOu6BBx7Ax8fH\nK/LuGXsD9mUjGPfE65Rs/gGly93MmIaTYGvAduQItmNHqf1kNVGF9egsKnxtoPCRiUyuoz7Wybf9\nJO6IGkj4bekcawfqsalsUXdl3CB3Hl3lijUoj9mZ+IOS1GEvQFwfUiNTCdIEkVdfgN5PAoUKuo34\nrR/T2TmD6ykQCASCvxbC0RNcVORU5FA2dwrt99cAnD7tomllKsCmBXx0pAN3zH0Zh6c5eGKHOHJn\npuKLEY5swVRsoo3BhVEDCq0Dy9EyXNEaLL4y2roGAPz796ft0CHUr1uPbtw15JevYHFREcO2OvlZ\n8T2ZESayxqtIN/+EfuD7bmeuaDs63Y/QVoWu80lyvn0UfZtoMmoMZPS4Cyy1ZFTtAJcDMudAm8t+\nv8bEoqJWIBAILgmE0BNcVOhz9VT0qmMSoVztKa5oRpOQZKXBwuUzvuaEu60dfgqJTWveok/tF3Ak\nk5zYFPQdujLxCweBpQYAigNVRHWPYtVAf/bUFTBpm5bojDF8evhTxq5dTcdHZqBNS0O/bjPDPjxG\nWgGkqEKRNcG07V5E9yAj47+5B3QJPF5vJzWkjviBnmuXDGRV14DJjD57Ffq7VruF6Qe3grkavpkJ\nvsFiEoVAIBAIfjOE0BNcNHz800JyS7JoE60i6rmpaJPSTtsnJ20selcJe/SH+fGHPO/22zuGEXNn\nOJqaj+CaOeTIZqYpa6h1mDD3lZiWD/42CK2H566vwOAvcTJA4pnRZqKq3uTWT4tQFsB3M8djHxdK\nmVbNF+kKtC4nsQ3l+Jc56GZSsWBcKDalAuoK0CuDyfD1RR8cSEatgQw5GMKiybCXgqEEFvcFTQAM\nfgb2fw4Wg8ibEwhawW63U1xcjMUza1oguJTw9fUlNjYWtVr9q44XQk9w0fDqwfdpkEDlspGavQpa\nmSIxefHr/Pif7TT2S/FL8KPd4/HkqdXkAfWhwejj+qCPakdtaTEhPiGMSxxJafibRFe5xd7NW53M\nv+3UR6PCdII1/RWAi1X9lZyUaqm1KSFOwd7xCnocdzDsB1jVX4lN6R6LFuB0klF2GH1wIFlaP9CG\n81LbDGZ9vBFdJy1Yc09dtG+wW9i1DDsLBAIAiouLCQwMpH379kiS9EdfjkDwuyHLMlVVVRQXF9Oh\nQ4dfdQ4h9AQXDU8oh2J+/yv8rgyAe5uLob1799KvXz/q691zyYLDgnnoxWv5gn0M84vjuOUkBEWT\n0dd9XEZKBm0LDIzNdOFv2Y6lDIrDwWmC/AQFnZwSZlmmTgl/73onBOfxUtQuIl0yD8khfOrrj6m+\nFK3Tzo3xMbwXW00ZLmIcEOa087gzgNSe48jIew8i2pMx4DkqZy3GuHUrWC4n/upeYGtwO3qNwk7k\nzQkErWKxWITIE1ySSJJEeHg4J0+e/NXnEEJPcNGQvMmAsUBGUaPEdJMabRw4HA569OhBXp47TKtU\nKnkxPZ1Jr7yCNi2NuWc4V2pkKmHbAzBu3wrJyfj3709nz2i0iaZ+xE9cetox487wfWs/e9cZOge9\n53uTJ6dQN2UypJ0edhYIBGdGiDzBpcr/+t++aK8iuGjQTZmMIiQEV20tlYuXMG3aNNRqtVfk3X77\n7RydcC/XlldQuXjJWc9lys7GWVeHb3IybWbOIP7NpbSZOQP//v29EzTOduyvae/S2LBZK0SeQPCX\nZO3atezxzNEWCP4sCKEnuGjQpqUR98YStsXHk7D8XV5//XUAOnbsSH19PR988AG6KZPPSaxVLl6C\nZc8elEFBaNPSMGVnUz5/Ac66ul+8jsaZub8kJgUCwV8HpVJJamoqSUlJ3HTTTdTW1jZ7/ZtvvuH7\n77/n8ssv/1Xnf+edd5g6dSoAa9asYd++fb94zLnud6kycOBAdu7c+buu+eSTTxIXF0dAQECz7Var\nlVtvvZXOnTvTt29fjh079rtdkxB6gouG0o0b6di/P/eu+w6b3Y6fnx9btmzhyJEj3g9VU9es/IUX\n2J90OeUvvHCaC9dSEDYKP8uePV4Bdybn7lzFpEAg+Ovg5+dHTk4Oe/fuJSwsjMWLFzd7ffjw4Sxc\nuPA3CTFfbELP4XD80Zfwp+Gmm25i+/bTR0u+9dZbhIaGcvjwYR555BGeeOKJ3+2ahNATXBTceOON\nxAwaRLnJhATMmjULk8lE//79z3hM9TvvgsNB9TvvnubCtQyj6qZMxjc5Gd/k5GbirzXnToRgBYJL\nmyuvvJKSkhLvz88//zy9e/cmOTmZZ555BgCj0cgNN9xASkoKSUlJrFy5EoD27dtTWekee7hz504G\nDhzY7NwwjF+XAAAgAElEQVQ//PADn3/+OY899hipqakcOXKEpUuX0rt3b1JSUhgzZgwmk6nV/V59\n9VUuu+wykpOTue2220677nfeeYebb76ZgQMHkpCQwJw5c7yvvfTSSyQlJZGUlMQrr7wCwLFjx0hK\nSvLu88ILL/Dss88Cbrfs4YcfplevXixatKjZOs8++ywTJ05k4MCBdOzYkVdfffWsz2Tu3Ln07t2b\npKQkHnjgAWRPc/uBAwfyyCOP0KtXL7p168aOHTsYPXo0CQkJzJ4923uNXbt25Y477qBbt26MHTsW\nk8l02r1/9913XHnllfTo0YNx48bR0OBuhj9jxgzvM3v00UfP+J6fK1dccQVRUVGnbf/ss8+45557\nABg7diyZmZne+7zQiGIMwZ+aJUvcuXgulwuAtIgINq1dS1Dv3r94bNiEe6h+513CJtxD4ODBAGd0\n4bRpaXT4aGWzbbqmxRMCgUAAOJ1OMjMzue+++wC3gDh06BDbt29HlmVGjBjB5s2bOXnyJNHR0axd\nuxYAg8FwTufv168fI0aM4MYbb2Ts2LEAhISEMGnSJABmz57NW2+9xbRp007bb8GCBRQUFODj43Na\naLmR7du3s3fvXrRaLb179+aGG25AkiTefvtttm3bhizL9O3bl6uvvprQ0NCzXqvNZjtjaPTAgQNs\n3LiR+vp6unTpwoMPPsg333zz/+3de1hVZd7w8e8NKAiaiXgKmUpFRdkIxEEkJrQYfRRhAhpthMnY\n2puHDr4jY16DoOY8kx0oeT105WD4jBj1kGIXzGMeYkxNRcQDhKhUKtKMKCaBIIqu94+92Q/ExiO6\nEX6f61oXe6/DvX7rXiz5ea97rdtsncyePZuEhAQAYmJiyMrKYuLEiQB07tyZvLw8li1bRnh4OAcO\nHMDR0ZGBAwcyZ84cAI4dO0ZKSgqBgYHExsaycuXKJknb+fPnWbJkCdu2bcPBwYGlS5eSlJTErFmz\n2LhxI8XFxSilzNZZTk6OaT+N2dvb880339ywfhorKyvDxcUFABsbG7p3705FRQVOTk63XMadkkRP\ntEnHjh3Dz8+Pn4195nr06EFeXh4DBgy45TL6zJ1Ln0YXe7Ph0m6ioeVOCPHgOXDqJ5ZtO85rzwzm\niUdvnLDcitraWjw9PSkrK8PNzY2QkBDAkOht2bIFL2MLf3V1NSdOnCAoKIg//vGPzJs3j9DQUIKC\ngu5434WFhcTHx3Px4kWqq6sZO3as2fU8PDyYMmUKv/3tb/ntb39rdp2QkBB69uwJQEREBLt27UIp\nxbPPPouDg4Np/s6dOwkLu/EY3JMmTWpx2YQJE7C1tcXW1pbevXtz9uxZdDqd2TrJycnh7bffpqam\nhgsXLjB8+HBTotcQg06nY/jw4abWsgEDBlBaWsrDDz+Mi4sLgYGBAERHR5OcnNwk0du7dy9FRUWm\nda5cuUJAQADdu3fHzs4OvV5PaGgooaGhzY5j9OjRHDp06Ib10NbJrVvRpjS8LmXo0KH8/PPPWFtb\ns2bNGi5cuHBbSZ4QomNbtu04X584z7Jtx1ulvIY+eqdOnULTNFMfPU3TmD9/PocOHeLQoUOUlJSg\n1+sZPHgw+fn56HQ64uPjWbzY8LInGxsb0x2KWx3pY+rUqSxfvpyCggISExNb3C47O5tZs2aRn5+P\nr6+v2b5zv+xDeKM+hY1jNRdvQ2Jojq2tremztbU19fX1Zuvk8uXLzJw5k4yMDAoKCpg+fXqT/TSU\nY2Vl1aRMKysr0/Hd7Jg0TSMkJMR0joqKikhJScHGxobc3FyioqLIyspi3LhxzY4jJycHT0/PZtOo\nUaNaPHZznJ2dKS0tBQx/5yorK00J970miZ5oM+bOnUunTp04aHz4ITIykvr6el588UULRyaEeNC8\n9sxgfu3qxGvPDG7Vcu3t7UlOTua9996jvr6esWPHsmbNGlOfr7KyMsrLy/nxxx+xt7cnOjqauLg4\n8vPzAUMfvQMHDgDw+eefm91Ht27dTC9/B6iqqqJfv35cvXqVtLQ0s+tdv36d0tJSRo8ezdKlS6ms\nrDTF1NjWrVu5cOECtbW1ZGZmEhgYSFBQEJmZmdTU1HDp0iU2btxIUFAQffr0oby8nIqKCurq6sjK\nyrqrujNXJw1JnZOTE9XV1WRkZNx2uadPn2bPnj0ArF+/vlnf7ZEjR7J7925KSkoAQ1/B48ePU11d\nTWVlJePHj+f999/n8OHDzcpuaNH75XQ7t23B0DK5du1aADIyMhgzZsx9ezek3LoVFrdt2zZCQ0Op\nq6sD4NFHH6WwsLDZ4+lCCHGrnni0B/+l978nZXt5eeHh4cEnn3xCTEwMR48eJSAgAICuXbuybt06\nSkpKiIuLw8rKik6dOrFq1SoAEhMT0ev1LFiwoNmDGA0mT57M9OnTSU5OJiMjgzfffBN/f3969eqF\nv7+/KblrvF56ejp6vZ7Kyko0TePVV1/l4Ycfbla2n58fkZGRnDlzhujoaHx8fABDq6Gfnx8A06ZN\nM92KTkhIwM/PD2dnZ4YOHXpX9VZQUNCsThr6H7q7u9O3b198b6H/9S8NGTKEFStWEBsby7Bhw5gx\nY0aT5b169SI1NZXnn3/e9HdmyZIldOvWjfDwcC5fvoymaSQlJd3V8QH86U9/Yv369dTU1NC/f3+m\nTZvGwoUL0ev1xMTEMGjQIBwdHUlPT7/rfd0qdb+e+mjrfHx8tPv9vp2OrqqqCnd3d06fPg0YBm7O\nzs5mzJgxFo5MCNGWHD16FDc3N0uH8cBLTU0lLy/P9A7S9uDkyZOEhoZSWFho6VDuKXPXgFLqgKZp\nPjfbtt3eulVKjVNKHVNKlSilZJT4NiYyMpKHHnqI06dPo5Ri7ty51NbWSpInhBBCtKJ2eetWKWUN\nrABCgDPAfqXUF5qmWf7Nkh3c3/72N15++WWuXbsGwBNPPMHevXuxsWmXv4pCCNFmTJ06lalTp1o6\njFb12GOPtfvWvLvVXv+6+gElmqZ9D6CUSgfCAUn0LKSkpARfX1/Te4q6d+/Ovn37GDJkiIUjE0II\nIdqv9nrr1hkobfT9jHGeuM+uXbtGQEAArq6uXLx4EWtra1atWsXFixclyRNCCCHusfaa6N0SpdRL\nSqk8pVTeuXPnLB1Ou5OcnIydnR179+4FDI+X19fX8/LLL1s4MiGEEKJjaK+3bssAl0bf+xvnNaFp\n2kfAR2B46vb+hNb+HTlyhKeffto0nqOXlxdfffWV2Uf9hRCivcjOzsbFxQUPDw9LhyKESXtt0dsP\nuCqlHldKdQYmA19YOKZ2r7a2lpEjRzJixAjOnz9Pjx492Lt3L/n5+ZLkCSEeaNbW1nh6euLu7s7E\niRObjYu6efNmduzYgU6nu6PyU1NTmT17NgCZmZkUFd28S/mtrtdRBQcHtzgW773c55AhQ0wjaJSX\nlwNQV1fHpEmTGDRoEP7+/pw8efK+xdQuEz1N0+qB2cCXwFHgM03TvrVsVO3bH//4RxwcHNi3bx/W\n1ta89dZbXLhwAX//e/PCUiGEuJ8ahkArLCzE0dHRNARag3HjxvH222+3ymgHD1qiZ26otY4sLS3N\nNIJG7969AUhJSaFHjx6UlJQwZ84c5s2bd9/iaZeJHoCmaf/QNG2wpmkDNU37i6Xjaa+ys7Pp2rUr\nSUlJaJrGuHHjqKuru6+/xEIIcT8FBARQVva/vYHeeecdfH198fDwIDExETAMszVhwgRGjBiBu7s7\nn376KWB4HUhDt5a8vLxmo2N88803fPHFF8TFxeHp6cl3333H6tWr8fX1ZcSIEURGRlJTU2N2veTk\nZIYNG4aHhweTJ09uFndqairh4eEEBwfj6urKokWLTMuSkpJwd3fH3d2dDz74ADC8jNjd3d20zrvv\nvsvChQsBQ8vV66+/jo+PD8uWLWuyn4ULFxIbG0twcDADBgwgOTn5hnWyePFifH19cXd356WXXqJh\nIIfg4GDmzJmDj48Pbm5u7N+/n4iICFxdXYmPjzfFOHToUKZMmYKbmxtRUVHU1NQ0O/YtW7YQEBCA\nt7c3zz33nGl4uDfeeMNUZ3Pnzm3xnN+tTZs28cILLwAQFRXF9u3buV8DVrTXPnriHvv3v/9NUFCQ\naexAFxcXduzYweOPP27hyIQQ4t65du0a27dvR6/XA4YE4sSJE+Tm5qJpGmFhYXz99decO3eORx55\nhOzsbAAqKytvqfxRo0YRFhZGaGgoUVFRAKZhwgDi4+NJSUnhlVdeabbeW2+9xQ8//ICtrW2zW8sN\ncnNzKSwsxN7eHl9fXyZMmIBSio8//ph9+/ahaRr+/v489dRT9OjR44axXrlypcVbo8XFxeTk5FBV\nVcWQIUOYMWMGmzdvNlsns2fPJiEhAYCYmBiysrKYOHEiAJ07dyYvL49ly5YRHh7OgQMHcHR0ZODA\ngcyZMweAY8eOkZKSQmBgILGxsaxcubJJ0nb+/HmWLFnCtm3bcHBwYOnSpSQlJTFr1iw2btxIcXEx\nSimzdZaTk2PaT2P29vYtjnf74osvYm1tTWRkJPHx8SilKCsrw8XF8OiAjY0N3bt3p6KiAicnpxvW\ncWtoty164t64du0azz33HP369aOkpAQ7OzvS09M5ffq0JHlCiLajNBf+HmH42Qpqa2vx9PSkb9++\nnD17lpCQEMCQ6G3ZsgUvLy+8vb0pLi7mxIkT6HQ6tm7dyrx589i5cyfdu3e/430XFhYSFBSETqcj\nLS2Nb7813xPJw8ODKVOmsG7duhZfQh8SEkLPnj3p0qULERER7Nq1i127dvHss8/i4OBA165diYiI\nYOfOnTeNa9KkSS0umzBhAra2tjg5OdG7d2/Onj3bYp3k5OTg7++PTqfjq6++anJ8YWFhAOh0OoYP\nH06/fv2wtbVlwIABlJYa3qLm4uJCYGAgANHR0ezatatJLHv37qWoqIjAwEA8PT1Zu3Ytp06donv3\n7tjZ2aHX69mwYQP29vbNjmP06NGm27CNp5aSvLS0NAoKCti5cyc7d+7k73//+03r8V6TRE/csg8/\n/JAuXbqQkZGBUopp06ZRW1t7w4tdCCEs4p9vwXfbDT9bQUMfvVOnTqFpmqmPnqZpzJ8/35QAlJSU\noNfrGTx4MPn5+eh0OuLj41m8eDFgaM25fv06AJcvX76lfU+dOpXly5dTUFBAYmJii9tlZ2cza9Ys\n8vPz8fX1Ndt37pd9CG/Up7BxrObidXBwaHFbW1tb02dra2vq6+vN1snly5eZOXMmGRkZFBQUMH36\n9Cb7aSjHysqqSZlWVlam47vZMWmaRkhIiOkcFRUVkZKSgo2NDbm5uURFRZGVlcW4ceOaHUdOTo7p\nwYrG06hRo8wet7Oz4ZW93bp14/e//z25ubmm+Q2JaX19PZWVlfTs2bPF+mtNkuiJmyosLKRPnz7M\nmDGDq1evotPp+Omnn1i9erWlQxNCCPOC34CBTxt+tiJ7e3uSk5N57733qK+vZ+zYsaxZs8bU56us\nrIzy8nJ+/PFH7O3tiY6OJi4ujvz8fMDQR+/AgQMAfP7552b30a1bN6qqqkzfq6qq6NevH1evXiUt\nLc3setevX6e0tJTRo0ezdOlSKisrTTE1tnXrVi5cuEBtbS2ZmZkEBgYSFBREZmYmNTU1XLp0iY0b\nNxIUFESfPn0oLy+noqKCuro6srKy7qruzNVJQ1Ln5OREdXU1GRkZt13u6dOn2bNnDwDr16/nySef\nbLJ85MiR7N6929TV6NKlSxw/fpzq6moqKysZP34877//PocPH25W9u206NXX15v6X169epWsrCxT\nH8ewsDDWrl0LQEZGBmPGjGmVB3duhfTREy26cuUKY8aMYffu3YBh2LLs7GxTE7kQQrRZLn4Qs+Ge\nFO3l5YWHhweffPIJMTExHD16lICAAAC6du3KunXrKCkpIS4uDisrKzp16sSqVasASExMRK/Xs2DB\ngmYPYjSYPHky06dPJzk5mYyMDN588038/f3p1asX/v7+puSu8Xrp6eno9XoqKyvRNI1XX33V7Gut\n/Pz8iIyM5MyZM0RHR+Pj4wMYWg39/PwAmDZtGl5eXgAkJCTg5+eHs7MzQ4cOvat6KygoaFYnDf0P\n3d3d6du3L76+vrdd7pAhQ1ixYgWxsbEMGzaMGTNmNFneq1cvUlNTef7556mrqwNgyZIldOvWjfDw\ncC5fvoymaSQlJd3V8dXV1TF27FiuXr3KtWvXeOaZZ0x9K/V6PTExMQwaNAhHR0fS09Pval+3Q92v\npz7aOh8fH+1+v2+nLZs/fz5vv/02169fx8rKikWLFpmechJCiPvp6NGjuLm5WTqMB15qaip5eXks\nX77c0qG0mpMnTxIaGkphYaGlQ7mnzF0DSqkDmqb53GxbadETTRw8eJCoqCi+//57AJ5++mm+/PJL\nrK2tLRyZEEIIIW6X9NETAPz888+MHz8eb29vvv/+e/z9/SkpKWHbtm2S5AkhRDvQ8FBHe/LYY4+1\n+9a8uyWJnmDhwoU4OTnxP//zP/Tt25dt27axd+9eBg4caOnQhBBCCHEXJNHrwL788kt69+7NokWL\nUErx5ptv8q9//Yunn37a0qEJIYQQohVIH70OqKysjIiICNP7fcLCwkhLS6Nr164WjkwIIYQQrUla\n9DqQ+vp6XnrpJX71q1+Rm5uLq6srR44cYdOmTZLkCSGEEO2QJHodxNq1a+nRowerV6+mS5curFmz\nhuPHj6PT6SwdmhBCtAvZ2dkcOXLE0mEI0YQkeu1cYWEhgwcPZurUqdTU1DB9+nQuXrzIiy++aOnQ\nhBDigWFtbY2npyfu7u5MnDiRixcvNlm+efNmduzYccf/eU5NTWX27NkAZGZmUlRUdNNtbnW9jio4\nOJj7/X7cP//5z7i4uDS7S1ZXV8ekSZMYNGgQ/v7+nDx50rTsr3/9K4MGDWLIkCF8+eWXrR6TJHrt\n1KVLlwgPD0en03HixAl8fX05deoUH330UYuDXQshhDCvYazbwsJCHB0dTWPdNhg3bhxvv/12qwxr\n9aAleubG1O2oJk6caOr/3lhKSgo9evSgpKSEOXPmMG/ePACKiopIT0/n22+/ZfPmzcycOZNr1661\nakyS6LVDf/nLX3B0dOSLL76gV69ebN68mdzcXPr372/p0IQQ4oEXEBBAWVmZ6fs777yDr68vHh4e\nJCYmAob/bE+YMIERI0bg7u7Op59+Chje+9YwHmpeXl6zYdC++eYbvvjiC+Li4vD09OS7775j9erV\n+Pr6MmLECCIjI6mpqTG7XnJyMsOGDcPDw4PJkyc3izs1NZXw8HCCg4NxdXVl0aJFpmVJSUm4u7vj\n7u7OBx98ABhGnWgYqxXg3XffZeHChYChtez111/Hx8eHZcuWNdnPwoULiY2NJTg4mAEDBpCcnHzD\nOlm8eDG+vr64u7vz0ksv0TBiV3BwMHPmzMHHxwc3Nzf2799PREQErq6uppGaTp48ydChQ5kyZQpu\nbm5ERUVRU1PT7Ni3bNlCQEAA3t7ePPfcc6ZxgN944w1Tnc2dO7fFc36rRo4cSb9+/ZrN37RpEy+8\n8AIAUVFRbN++HU3T2LRpE5MnT8bW1pbHH3+cQYMGmU0U74Y07bQj27dvZ8qUKZw9e5ZOnTqRmJho\nuiiFEELcvWvXrrF9+3b0ej1gSCBOnDhBbm4umqYRFhbG119/zblz53jkkUfIzs4GoLKy8pbKHzVq\nFGFhYYSGhhIVFQVgGg8WID4+npSUFF555ZVm67311lv88MMP2NraNru13CA3N5fCwkLs7e3x9fVl\nwoQJKKX4+OOP2bdvH5qm4e/vz1NPPUWPHj1uGOuVK1davDVaXFxMTk4OVVVVDBkyhBkzZrB582az\ndTJ79mwSEhIAiImJISsri4kTJwLQuXNn8vLyWLZsGeHh4Rw4cABHR0cGDhzInDlzADh27BgpKSkE\nBgYSGxvLypUrmyRt58+fZ8mSJWzbtg0HBweWLl1KUlISs2bNYuPGjRQXF6OUMltnOTk5pv00Zm9v\nzzfffHPD+mmsrKwMFxcXAGxsbOjevTsVFRWUlZUxcuRI03r9+/dv8p+I1iAteu3Av//9b0aNGsUz\nzzzD2bNn+Y//+A/Onz8vSZ4QosM6VH6Il7e+zKHyQ61SXm1tLZ6envTt25ezZ88SEhICGBK9LVu2\n4OXlhbe3N8XFxZw4cQKdTsfWrVuZN28eO3fupHv37ne878LCQoKCgtDpdKSlpfHtt9+aXc/Dw4Mp\nU6awbt26FrvohISE0LNnT7p06UJERAS7du1i165dPPvsszg4ONC1a1ciIiLYuXPnTeOaNGlSi8sm\nTJiAra0tTk5O9O7dm7Nnz7ZYJzk5Ofj7+6PT6fjqq6+aHF9YWBgAOp2O4cOH069fP2xtbRkwYACl\npaUAuLi4EBgYCEB0dDS7du1qEsvevXspKioiMDAQT09P1q5dy6lTp+jevTt2dnbo9Xo2bNiAvb19\ns+MYPXo0hw4dajbdTpJnaZLoPcA0TWP9+vUMGjSIPXv2MGDAAPLz8/nHP/7BQw89ZOnwhBDCYj48\n/CG7f9zNh4c/bJXyGvronTp1Ck3TTH30NE1j/vz5pgSgpKQEvV7P4MGDyc/PR6fTER8fz+LFiwFD\na87169cBuHz58i3tu2HosoKCAhITE1vcLjs7m1mzZpGfn4+vr6/ZvnO/7EN4oz6FjWM1F6+Dg0OL\n29ra2po+W1tbU19fb7ZOLl++zMyZM8nIyKCgoIDp06c32U9DOVZWVk3KtLKyMh3fzY5J0zRCQkJM\n56ioqIiUlBRsbGzIzc0lKiqKrKwsxo0b1+w4cnJy8PT0bDaNGjWqxWM3x9nZ2ZSY1tfXU1lZSc+e\nPZvMBzhz5gzOzs63VfbNSKL3gDp48CBBQUFMmTKFQYMG8dFHH/Hdd9/h5eVl6dCEEMLiXh7xMoGP\nBPLyiJdbtVx7e3uSk5N57733qK+vZ+zYsaxZs8bU56usrIzy8nJ+/PFH7O3tiY6OJi4ujvz8fMDQ\nR+/AgQMAfP7552b30a1bN6qqqkzfq6qq6NevH1evXiUtLc3setevX6e0tJTRo0ezdOlSKisrTTE1\ntnXrVi5cuEBtbS2ZmZkEBgYSFBREZmYmNTU1XLp0iY0bNxIUFESfPn0oLy+noqKCuro6srKy7qru\nzNVJQ1Ln5OREdXU1GRkZt13u6dOn2bNnDwDr16/nySefbLJ85MiR7N69m5KSEsDQV/D48eNUV1dT\nWVnJ+PHjef/99zl8+HCzslurRS8sLIy1a9cCkJGRwZgxY1BKERYWRnp6OnV1dfzwww+cOHECPz+/\n266DG5E+eg+Yc+fOER8fz+rVq+nVqxcpKSlMnToVKyvJ2YUQooFnb08+DGmd1rxf8vLywsPDg08+\n+YSYmBiOHj1KQEAAAF27dmXdunWUlJQQFxeHlZUVnTp1YtWqVQAkJiai1+tZsGBBswcxGkyePJnp\n06eTnJxMRkYGb775Jv7+/vTq1Qt/f39Tctd4vfT0dPR6PZWVlWiaxquvvsrDDz/crGw/Pz8iIyM5\nc+YM0dHR+Pj4AIZWw4YEY9q0aaZGg4SEBPz8/HB2dmbo0KF3VW8FBQXN6qSh/6G7uzt9+/bF19f3\ntssdMmQIK1asIDY2lmHDhjFjxowmy3v16kVqairPP/88dXV1ACxZsoRu3boRHh7O5cuX0TSNpKSk\nuzo+gD/96U+sX7+empoa+vfvz7Rp01i4cCF6vZ6YmBgGDRqEo6Mj6enpAAwfPpzf/e53DBs2DBsb\nG1asWIG1tfVdx9GYani6paPz8fHR7vf7dm7H1atXWbVqFYmJiVRXV/Pqq6+SkJBwV/0+hBDiQXD0\n6FHc3NwsHcYDLzU1lby8PJYvX27pUFrNyZMnCQ0NpbCw0NKh3FPmrgGl1AFN03xutq206D0Atm/f\nzmuvvca3337Lb37zGz744AP5R08IIYQQNyX3+9qwkydPEhkZyTPPPENtbS2bNm1i8+bNkuQJIYS4\nbQ0PdbQnjz32WLtvzbtb0qLXBtXU1PDWW2/xzjvvYGVlxX/+538yZ84c7OzsLB2aEEIIIR4gkui1\nIZqm8dlnnxEXF0dpaSm///3vWbp0qYxoIYQQQog7Irdu24jDhw8THBzM5MmTcXJyYufOnaSlpUmS\nJ4QQQog7JomehVVUVDBz5ky8vb0pKirio48+Yv/+/c3eAySEEKJty87O5siRI5YOQ4gmJNGzkPr6\nelauXImrqysfffQRs2fP5vjx40yfPr3V36EjhBDi7lhbW+Pp6Ym7uzsTJ05sNi7q5s2b2bFjBzqd\n7o7KT01NZfbs2QBkZmZSVFR0021udb2OKjg4uMWxeO+FqqqqJqNnODk58frrrwNw6tQpnn76aTw8\nPAgODubMmTOm7dauXYurqyuurq6mlyq3Jkn0LOCf//wn3t7ezJo1C29vbw4fPsyyZctuOoC0EEII\ny2gYAq2wsBBHR0fTEGgNxo0bx9tvv33DIcVu1YOW6Jkbaq0j6tatW5PRMx599FEiIiIAmDt3Ln/4\nwx84cuQICQkJzJ8/H4ALFy6waNEi9u3bR25uLosWLeKnn35q1bgk0buPTp8+ze9+9ztGjx5NVVUV\nn3/+OVu3bmX48OGWDk0IIcQtCggIoKyszPT9nXfewdfXFw8PDxITEwHDMFsTJkxgxIgRuLu78+mn\nnwKG14GcP38egLy8vGajY3zzzTd88cUXxMXF4enpyXfffcfq1avx9fVlxIgRREZGUlNTY3a95ORk\nhg0bhoeHB5MnT24Wd2pqKuHh4QQHB+Pq6sqiRYtMy5KSknB3d8fd3Z0PPvgAMLziy93d3bTOu+++\ny8KFCwFDa9nrr7+Oj48Py5Yta7KfhQsXEhsbS3BwMAMGDCA5OfmGdbJ48WJ8fX1xd3fnpZdeomEg\nh+DgYObMmYOPjw9ubm7s37+fiIgIXF1diY+PN8U4dOhQpkyZgpubG1FRUdTU1DQ79i1bthAQEIC3\nt0yvYpQAAA6WSURBVDfPPfecaXi4N954w1Rnc+fObfGc367jx49TXl5OUFAQAEVFRYwZMwYwDKu2\nadMmAL788ktCQkJwdHSkR48ehISEsHnz5laLA+Sp2/smOTmZN954AzD8Us+dO5cuXbpYOCohhBC3\n49q1a2zfvh29Xg8YEogTJ06Qm5uLpmmEhYXx9ddfc+7cOR555BGys7MBqKysvKXyR40aRVhYGKGh\noURFRQGYhgkDiI+PJyUlhVdeeaXZem+99RY//PADtra2zW4tN8jNzaWwsBB7e3t8fX2ZMGECSik+\n/vhj9u3bh6Zp+Pv789RTT930LtOVK1davDVaXFxMTk4OVVVVDBkyhBkzZrB582azdTJ79mwSEhIA\niImJISsri4kTJwLQuXNn8vLyWLZsGeHh4Rw4cABHR0cGDhzInDlzADh27BgpKSkEBgYSGxvLypUr\nmyRt58+fZ8mSJWzbtg0HBweWLl1KUlISs2bNYuPGjRQXF6OUMltnOTk5pv00Zm9vf8PxbtPT05k0\naZKphXfEiBFs2LCB1157jY0bN1JVVUVFRQVlZWW4uLiYtuvfv3+T/0S0BmnRu08eeughwsLCKC4u\nZsGCBZLkCSHEPVRz8CCnp02n5uDBVimvtrYWT09P+vbty9mzZwkJCQEMid6WLVvw8vLC29ub4uJi\nTpw4gU6nY+vWrcybN4+dO3fe1XCVhYWFBAUFodPpSEtL49tvvzW7noeHB1OmTGHdunXY2JhvxwkJ\nCaFnz5506dKFiIgIdu3axa5du3j22WdxcHCga9euREREsHPnzpvGNWnSpBaXTZgwAVtbW5ycnOjd\nuzdnz55tsU5ycnLw9/dHp9Px1VdfNTm+sLAwAHQ6HcOHD6dfv37Y2toyYMAASktLAXBxcSEwMBCA\n6Ohodu3a1SSWvXv3UlRURGBgIJ6enqxdu5ZTp07RvXt37Ozs0Ov1bNiwAXt7+2bHMXr06Ca3Yxum\nGyV5YEj0nn/+edP3d999lx07duDl5cWOHTtwdna+b/3xJdG7T6ZOnUp6ejq/+tWvLB2KEEK0e+dX\nrOTSrl2cX7GyVcpr6KN36tQpNE0z9dHTNI358+ebEoCSkhL0ej2DBw8mPz8fnU5HfHw8ixcvBsDG\nxobr168DcPny5Vvad8OIFgUFBSQmJra4XXZ2NrNmzSI/Px9fX1+zfed+2YfwRn0KG8dqLl4HB4cW\nt7W1tTV9tra2pr6+3mydXL58mZkzZ5KRkUFBQQHTp09vsp+GcqysrJqUaWVlZTq+mx2TpmmEhISY\nzlFRUREpKSnY2NiQm5tLVFQUWVlZjBs3rtlx5OTkNHnAomEaNWpUi8d++PBh6uvreeKJJ0zzHnnk\nETZs2MDBgwf5y1/+Ahhaap2dnU0JK8CZM2dwdnZusew7IYmeEEKIdsdp1kwcnnwSp1kzW7Vce3t7\nkpOTee+996ivr2fs2LGsWbPG1OerrKyM8vJyfvzxR+zt7YmOjiYuLo78/HzA0EfvwIEDAHz++edm\n99GtWzeqqqpM36uqqujXrx9Xr14lLS3N7HrXr1+ntLSU0aNHs3TpUiorK00xNbZ161YuXLhAbW0t\nmZmZBAYGEhQURGZmJjU1NVy6dImNGzcSFBREnz59KC8vp6Kigrq6OrKysu6q7szVSUNS5+TkRHV1\nNRkZGbdd7unTp9mzZw8A69evb/Z6spEjR7J7925KSkoAQ1/B48ePU11dTWVlJePHj+f999/n8OHD\nzcq+kxa9Tz75pElrHhhuHzckzX/961+JjY0FYOzYsWzZsoWffvqJn376iS1btjB27NjbroMbkT56\nQggh2h17Ly9+9bfV96RsLy8vPDw8+OSTT4iJieHo0aMEBAQA0LVrV9atW0dJSQlxcXFYWVnRqVMn\nVq1aBUBiYiJ6vZ4FCxY0exCjweTJk5k+fTrJyclkZGTw5ptv4u/vT69evfD39zcld43XS09PR6/X\nU1lZiaZpvPrqqzz88MPNyvbz8yMyMpIzZ84QHR2Nj48PYGg19PPzA2DatGl4eXkBkJCQgJ+fH87O\nzgwdOvSu6q2goKBZnTT0P3R3d6dv3774+vredrlDhgxhxYoVxMbGMmzYMGbMmNFkea9evUhNTeX5\n55+nrq4OgCVLltCtWzfCw8O5fPkymqaRlJR0V8fX4LPPPuMf//hHk3n//Oc/mT9/Pkopfv3rX5ta\nhB0dHVmwYIHpuBMSEnB0dGyVOBqohqdbOjofHx/tfr5vRwghxK05evQobm5ulg7jgZeamkpeXh7L\nly+3dCit5uTJk4SGhlJYWGjpUO4pc9eAUuqApmk+N9tWbt0KIYQQQrRTcutWCCGE6ACmTp3K1KlT\nLR1Gq3rsscfafWve3ZIWPSGEEEKIdkoSPSGEEG2e9CcXHdXd/u5LoieEEKJNs7Ozo6KiQpI90eFo\nmkZFRQV2dnZ3XIb00RNCCNGm9e/fnzNnznDu3DlLhyLEfWdnZ0f//v3veHtJ9IQQQrRpnTp14vHH\nH7d0GEI8kOTWrRBCCCFEOyWJnhBCCCFEOyWJnhBCCCFEOyVDoBkppc4BpywdRzvnBJy3dBAdnJyD\ntkHOg+XJOWgb5DzcuUc1Tet1s5Uk0RP3jVIq71bG5RP3jpyDtkHOg+XJOWgb5Dzce3LrVgghhBCi\nnZJETwghhBCinZJET9xPH1k6ACHnoI2Q82B5cg7aBjkP95j00RNCCCGEaKekRU8IIYQQop2SRE/c\nc0qpcUqpY0qpEqXUG5aO50GklHJRSuUopYqUUt8qpV4zzndUSm1VSp0w/uxhnK+UUsnGOj+ilPJu\nVNYLxvVPKKVeaDT/CaVUgXGbZKWUutE+OiqllLVS6qBSKsv4/XGl1D5jvX2qlOpsnG9r/F5iXP5Y\nozLmG+cfU0qNbTTf7LXS0j46KqXUw0qpDKVUsVLqqFIqQK6F+08pNcf471GhUuoTpZSdXA9tkKZp\nMsl0zybAGvgOGAB0Bg4Dwywd14M2Af0Ab+PnbsBxYBjwNvCGcf4bwFLj5/HA/wAKGAnsM853BL43\n/uxh/NzDuCzXuK4ybvsfxvlm99FRJ+D/AuuBLOP3z4DJxs8fAjOMn2cCHxo/TwY+NX4eZrwObIHH\njdeH9Y2ulZb20VEnYC0wzfi5M/CwXAv3/Rw4Az8AXYzfPwOmyvXQ9iZp0RP3mh9Qomna95qmXQHS\ngXALx/TA0TTtX5qm5Rs/VwFHMfxDG47hjx7Gn781fg4H/ksz2As8rJTqB4wFtmqadkHTtJ+ArcA4\n47KHNE3bqxn+9fyvX5Rlbh8djlKqPzAB+JvxuwLGABnGVX55DhrqLQN42rh+OJCuaVqdpmk/ACUY\nrhOz18pN9tHhKKW6A78GUgA0TbuiadpF5FqwBBugi1LKBrAH/oVcD22OJHriXnMGSht9P2OcJ+6Q\n8ZaHF7AP6KNp2r+Mi/4N9DF+bqnebzT/jJn53GAfHdEHwJ+A68bvPYGLmqbVG783rjdTXRuXVxrX\nv91zc6N9dESPA+eAj4230P+mlHJAroX7StO0MuBd4DSGBK8SOIBcD22OJHpCPECUUl2Bz4HXNU37\nufEyY+vDPX2M/n7so61SSoUC5ZqmHbB0LB2cDeANrNI0zQu4hOE2qolcC/eesX9iOIbE+xHAARhn\n0aCEWZLoiXutDHBp9L2/cZ64TUqpThiSvDRN0zYYZ5813mrC+LPcOL+ler/R/P5m5t9oHx1NIBCm\nlDqJ4TbSGGAZhluBNsZ1Gtebqa6Ny7sDFdz+uam4wT46ojPAGU3T9hm/Z2BI/ORauL+eAX7QNO2c\npmlXgQ0YrhG5HtoYSfTEvbYfcDU+JdUZQyfcLywc0wPH2C8lBTiqaVpSo0VfAA1PC74AbGo0/w/G\nJw5HApXGW05fAr9RSvUw/o/8N8CXxmU/K6VGGvf1h1+UZW4fHYqmafM1TeuvadpjGH6Pv9I0bQqQ\nA0QZV/vlOWiotyjj+ppx/mTjU4iPA64YOv+bvVaM27S0jw5H07R/A6VKqSHGWU8DRci1cL+dBkYq\npeyN9dRwHuR6aGss/TSITO1/wvDU23EMT1D92dLxPIgT8CSG20RHgEPGaTyG/irbgRPANsDRuL4C\nVhjrvADwaVRWLIYOzyXAi43m+wCFxm2W878vVDe7j448AcH871O3AzD8YSoB/huwNc63M34vMS4f\n0Gj7Pxvr+RjGJzqN881eKy3to6NOgCeQZ7weMjE8NSvXwv0/D4uAYmNd/R3Dk7NyPbSxSUbGEEII\nIYRop+TWrRBCCCFEOyWJnhBCCCFEOyWJnhBCCCFEOyWJnhBCCCFEOyWJnhBCCCFEOyWJnhBC3IRS\nqvo21l2olJp7r8oXQojbIYmeEEIIIUQ7JYmeEELcAaXURKXUPqXUQaXUNqVU4wHuRyil9iilTiil\npjfaJk4ptV8pdUQptchMmf2UUl8rpQ4ppQqVUkH35WCEEO2WJHpCCHFndgEjNU3zwjD27Z8aLfPA\nMBZuAJCglHpEKfUbDMM7+WEY2eEJpdSvf1Hm7zEMw+UJjMAwAooQQtwxm5uvIoQQwoz+wKfGwe07\nAz80WrZJ07RaoFYplYMhuXsSw3iqB43rdMWQ+H3daLv9wBqlVCcgU9M0SfSEEHdFWvSEEOLO/D9g\nuaZpOuD/YBjLs8Evx5bUMIy5+ldN0zyN0yBN01KarKRpXwO/BsqAVKXUH+5d+EKIjkASPSGEuDPd\nMSRkAC/8Ylm4UspOKdUTCMbQUvclEKuU6gqglHJWSvVuvJFS6lHgrKZpq4G/Ad73MH4hRAcgt26F\nEOLm7JVSZxp9TwIWAv+tlPoJ+Ap4vNHyI0AO4AS8qWnaj8CPSik3YI9SCqAaiAbKG20XDMQppa4a\nl0uLnhDirihN++UdBiGEEEII0R7IrVshhBBCiHZKEj0hhBBCiHZKEj0hhBBCiHZKEj0hhBBCiHZK\nEj0hhBBCiHZKEj0hhBBCiHZKEj0hhBBCiHZKEj0hhBBCiHbq/wPksrVMV08wHgAAAABJRU5ErkJg\ngg==\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbkAAAEWCAYAAAD7HukTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZfoAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "atnp5Lo6aaLe", "colab_type": "text" }, "source": [ "On constate que les performances diminuent (l'erreur augmente) avec le nombre d'échantillons d'entraînement, et que les performances en validation sont au départ beaucoup plus faibles, ce qui est signe d'**overfitting**. En effet, en utilisant seulement 50 exemples (gauche du graphe), on remarque que le modèle compte presque autant de paramètres (un par variable prédictive). Dans ce cas, les paramètres peuvent être adaptés aux exemples spécifiques d'entraînement pour apprendre leur prédiction presque \"par coeur\". Le modèle n'est alors pas générique et fonctionne mal sur de nouvelles données (comme le montrent les faibles performances en validation). On constate ensuite que pour 400 échantillons d'entraînement, les performances en entraînement se stabilisent, et les performances en validation sont équivalentes, ce qui signifie que le nombre d'échantillons est suffisant (plus d'overfitting)." ] }, { "cell_type": "code", "metadata": { "id": "K6tK_SqBZZY0", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "3f98d994-7f46-4d18-e47f-dcc4b422fb5e" }, "source": [ "len(price_predictor.coef_)" ], "execution_count": 400, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "28" ] }, "metadata": { "tags": [] }, "execution_count": 400 } ] }, { "cell_type": "markdown", "metadata": { "id": "JL63VJR3c7fu", "colab_type": "text" }, "source": [ "## Performances réelles\n", "\n", "Pour évaluer les performances réelles du modèles, nous devons utiliser un set de données qui n'a pas encore été utilisé, ni pour l'entraînement ni pour la validation. Cela permet de simuler une utilisation réelle du modèle, sur de nouvelles maisons (dont on voudrait estimer le prix). L'évaluation des performances à partir du set de validation seraient biaisée, car il a en effet permis de choisir les paramètres du modèles, en l'occurence le choix des variables prédictives (le modèle a donc été indirectement optimisé pour ces données spécifiques).\n", "\n", "Sur la plateforme [Kaggle](https://www.kaggle.com/), pour garantir l'absence de biais lors du design du modèle, un set de test est généralement fourni séparément, et les labels ne sont volontairement pas fournis. L'utilisateur de la plateforme doit alors soumettre à la plateforme les prédictions données par le modèle développé, et reçoit le résultat de l'évaluation effectuée par la plateforme. Cela permet notamment de classer différents compétiteurs lors d'un concours: https://www.kaggle.com/competitions\n", "\n", "Concernant le dataset utilisé, on peut soumettre des prédictions ici: https://www.kaggle.com/c/home-data-for-ml-course/overview/evaluation\n", "\n", "Remarque: lorsque le concours ne sera fini, il sera toujours possible de récupérer les données complète ici: https://www.kaggle.com/prevek18/ames-housing-dataset\n", "\n" ] }, { "cell_type": "code", "metadata": { "id": "i5eYYBMhTQJl", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "9313dd30-4059-49c0-ed2c-fd11bdf04f62" }, "source": [ "# Vérifions déjà les performances sur le set de validation\n", "#Régression linéaire\n", "\n", "X = ames[featurelist]\n", "y = ames[\"SalePrice\"]\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", "\n", "price_predictor.fit(Xtrain, ytrain)\n", "\n", "trainpreds = price_predictor.predict(Xtrain).clip(ytrain.min(), ytrain.max())\n", "valpreds_lin = price_predictor.predict(Xval).clip(ytrain.min(), ytrain.max())\n", "\n", "MAE(trainpreds,ytrain.values), MAE(valpreds_lin,yval.values)" ], "execution_count": 401, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(20164.0980898648, 21040.7804916666)" ] }, "metadata": { "tags": [] }, "execution_count": 401 } ] }, { "cell_type": "code", "metadata": { "id": "P_tHgbL2p_mz", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "632f07da-8ed1-4760-9826-1466a3f921a8" }, "source": [ "#Régression polynomiale\n", "X = ames[featurelist]\n", "X = add_polynomials(X)\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]\n", "\n", "poly_predictor = LinearRegression()\n", "poly_predictor.fit(Xtrain, ytrain)\n", "\n", "trainpreds = poly_predictor.predict(Xtrain).clip(ytrain.min(), ytrain.max())\n", "valpreds_poly = poly_predictor.predict(Xval).clip(ytrain.min(), ytrain.max())\n", "\n", "MAE(trainpreds,ytrain.values), MAE(valpreds_poly,yval.values)" ], "execution_count": 402, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(18891.879496570047, 19434.25335313196)" ] }, "metadata": { "tags": [] }, "execution_count": 402 } ] }, { "cell_type": "code", "metadata": { "id": "0xDnxJVEhBI0", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 439 }, "outputId": "b8986a91-c4e0-4608-87f2-846e139a38d9" }, "source": [ "display_results(yval.values, valpreds_lin, \"Régression linéaire\")\n", "display_results(yval.values, valpreds_poly, \"Régression polynomiale\")" ], "execution_count": 403, "outputs": [ { "output_type": "stream", "text": [ "Mean average error: 21040.7804916666\n", "Mean average error: 19434.25335313196\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAGDCAYAAAC8371AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noCeEEEIIUU1JoCeEEEIIUU39f87vtkJGirmtAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "2chtMCm7-T7k", "colab_type": "text" }, "source": [ "Remarquez que vous obtiendrez des résultats différents à chaque fois que vous relancez les cellules ci-dessus avec un random_state différent: cela est dû à la séparation aléatoire des sets d'entraînement et de validation. \n", "\n", "Afin d'avoir un indicateur plus robuste du modèle, une possibilité serait de calculer une moyenne de ces résultats, ou d'utiliser un processus de **k-fold cross-validation**: on découpe l'ensemble des données en k échantillons, et pour chaque échantillon on entraînement un modèle sur les autres données, et on calcule les prédictions pour l'échantillon gardé de côté. On calcule de cette manière des prédictions pour chaque échantillon et on extrait une mesure (e.g., MAE) sur l'ensemble des prédictions." ] }, { "cell_type": "markdown", "metadata": { "id": "NW14_M7l0N5X", "colab_type": "text" }, "source": [ "### Vérification du set de test" ] }, { "cell_type": "code", "metadata": { "id": "QM8eeCvzzsLa", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 114 }, "outputId": "e9f357aa-5656-4c2a-92d6-4cd68f1d7da6" }, "source": [ "testset = pd.read_csv(\"https://raw.githubusercontent.com/titsitits/Python_Data_Science/master/Donn%C3%A9es/test.csv\")\n", "\n", "ids = testset['Id']\n", "\n", "print(len(testset))\n", "\n", "#nettoyage du testset\n", "testset = clean_df(testset)\n", "\n", "#Colonnes incomplètes\n", "counts = testset.count()\n", "incomplete = counts[counts < len(testset)]\n", "display_series(incomplete.sort_values())\n", "len(incomplete.index)" ], "execution_count": 404, "outputs": [ { "output_type": "stream", "text": [ "1459\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
BsmtFullBathBsmtHalfBathBsmtFinSF1BsmtFinSF2BsmtUnfSFTotalBsmtSFGarageCarsGarageArea
014571457145814581458145814581458
\n", "
" ], "text/plain": [ " BsmtFullBath BsmtHalfBath BsmtFinSF1 ... TotalBsmtSF GarageCars GarageArea\n", "0 1457 1457 1458 ... 1458 1458 1458\n", "\n", "[1 rows x 8 columns]" ] }, "metadata": { "tags": [] } }, { "output_type": "execute_result", "data": { "text/plain": [ "8" ] }, "metadata": { "tags": [] }, "execution_count": 404 } ] }, { "cell_type": "markdown", "metadata": { "id": "J1KVfZDT4hkF", "colab_type": "text" }, "source": [ "Il reste deux observations contenant des variables invalides. On va simplement remplacer les valeurs manquantes par les moyennes. Si le nombre d'observations à nettoyer était plus conséquent, il serait pertinent d'analyser les variables à corriger." ] }, { "cell_type": "code", "metadata": { "id": "1BVOHMFc_H0D", "colab_type": "code", "colab": {} }, "source": [ "for col in incomplete.index:\n", " testset[col] = testset[col].fillna(ames[col].mean())" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "1ai3HrFnNLmt", "colab_type": "text" }, "source": [ "### Soumission des résultats" ] }, { "cell_type": "code", "metadata": { "id": "tSvjBcA7bEi9", "colab_type": "code", "colab": {} }, "source": [ "X = ames[featurelist]\n", "y = ames[\"SalePrice\"]\n", "#On peut ré-entraîner le modèle sur tout le dataset (entraînement+validation), pour le rendre potentiellement plus robuste\n", "price_predictor.fit(X, y)\n", "\n", "#select features\n", "Xtest = testset[featurelist]\n", "\n", "testpreds_lin = price_predictor.predict(Xtest).clip(ytrain.min(), ytrain.max())\n", "\n", "submission = ids.to_frame()\n", "submission[\"SalePrice\"] = testpreds_lin\n", "submission.to_csv(\"mysubmission_linear_regression.csv\", index = False)" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "DkM2iUzEpspz", "colab_type": "code", "colab": {} }, "source": [ "Xtest2 = add_polynomials(Xtest)\n", "\n", "X = ames[featurelist]\n", "X = add_polynomials(X)\n", "poly_predictor.fit(X, y)\n", "\n", "testpreds_poly = poly_predictor.predict(Xtest2).clip(ytrain.min(), ytrain.max())\n", "\n", "submission = ids.to_frame()\n", "submission[\"SalePrice\"] = testpreds_poly\n", "submission.to_csv(\"mysubmission_polynomial_regression.csv\", index = False)" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "23u9i5kTpKME", "colab_type": "text" }, "source": [ "## Exploration de base des variables catégorielles" ] }, { "cell_type": "code", "metadata": { "id": "KMD_aKxojyyV", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 331 }, "outputId": "3f90f7c3-0f87-498d-b368-20b4fe64999b" }, "source": [ "#Analysons l'effet de chaque variable catégorielle\n", "\n", "all_columns = ames.columns\n", "cat_columns = ames.columns[ames.dtypes == 'object']\n", "\n", "diffs_per_group = []\n", "\n", "#Bonus: vous pouvez afficher des graphes pour chaque groupe\n", "#ncols = 8\n", "#ncats = len(cat_columns)\n", "#fig, axes = plt.subplots(int(np.ceil(ncats/ncols)),ncols, figsize = (20,30))\n", "\n", "i=0\n", "for cat in cat_columns: \n", " \n", " group_means = ames.groupby(cat)[\"SalePrice\"].mean()\n", " diffs_per_group.append(group_means.max() - group_means.min())\n", " \n", " #Bonus: vous pouvez afficher des graphes pour chaque groupe\n", " #plt.figure(figsize=(10,3))\n", " #group_means.plot.bar()\n", " #ax = axes[int(i/ncols), int(i%ncols)]\n", " #ames.boxplot(\"SalePrice\", by=cat, ax = ax)\n", " i=i+1\n", "\n", "best_cats = pd.Series(diffs_per_group, list(cat_columns)).abs().sort_values(ascending = False)\n", "best_cats.plot.bar(figsize=(10,4))\n", "\n", "best_cats = best_cats.index.to_list()" ], "execution_count": 408, "outputs": [ { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmwAAAE6CAYAAACrqg5SAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRERERPQ5/h+0uHqSVF0B3QAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "XIoTezR8R3kr", "colab_type": "text" }, "source": [ "## Bonus - Un algorithme plus moderne: catboost" ] }, { "cell_type": "code", "metadata": { "id": "QdEdNo8uPa7F", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 277 }, "outputId": "993ea1c8-c4aa-444e-e301-c2268a5478c8" }, "source": [ "#Installer un nouveau package: on appelle une ligne de commande linux grâce au symbole \"!\". On utilise le gestionnaire de packages python pip pour installer un nouveau package\n", "!pip install catboost" ], "execution_count": 409, "outputs": [ { "output_type": "stream", "text": [ "Requirement already satisfied: catboost in /usr/local/lib/python3.6/dist-packages (0.17.3)\n", "Requirement already satisfied: plotly in /usr/local/lib/python3.6/dist-packages (from catboost) (4.1.1)\n", "Requirement already satisfied: pandas>=0.24.0 in /usr/local/lib/python3.6/dist-packages (from catboost) (0.24.2)\n", "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from catboost) (1.12.0)\n", "Requirement already satisfied: graphviz in /usr/local/lib/python3.6/dist-packages (from catboost) (0.10.1)\n", "Requirement already satisfied: matplotlib in /usr/local/lib/python3.6/dist-packages (from catboost) (3.0.3)\n", "Requirement already satisfied: numpy>=1.16.0 in /usr/local/lib/python3.6/dist-packages (from catboost) (1.16.5)\n", "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (from catboost) (1.3.1)\n", "Requirement already satisfied: retrying>=1.3.3 in /usr/local/lib/python3.6/dist-packages (from plotly->catboost) (1.3.3)\n", "Requirement already satisfied: pytz>=2011k in /usr/local/lib/python3.6/dist-packages (from pandas>=0.24.0->catboost) (2018.9)\n", "Requirement already satisfied: python-dateutil>=2.5.0 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.24.0->catboost) (2.5.3)\n", "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->catboost) (2.4.2)\n", "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->catboost) (1.1.0)\n", "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib->catboost) (0.10.0)\n", "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from kiwisolver>=1.0.1->matplotlib->catboost) (41.2.0)\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "id": "WDhe159xjnZ6", "colab_type": "code", "colab": {} }, "source": [ "contfeatures = best_features[:28]\n", "catfeatures = best_cats[:-5] #on retire les plus mauvaises\n", "\n", "newfeaturelist = contfeatures + catfeatures" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "wuMmjKHDQYqT", "colab_type": "code", "colab": {} }, "source": [ "X = ames[newfeaturelist]\n", "y = ames[\"SalePrice\"]\n", "\n", "Xtrain, Xval, ytrain, yval = X.loc[trainids], X.loc[valids], y.loc[trainids], y.loc[valids]" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "4oM1lPqJiVF0", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "382a4593-fdd5-41a3-c70e-3941fdc31365" }, "source": [ "import numpy as np\n", "from catboost import Pool, CatBoostRegressor\n", "\n", "\n", "cat_feature_ids = [i for i in range(len(newfeaturelist)) if newfeaturelist[i] in catfeatures]\n", "\n", "\n", "train_pool = Pool(Xtrain.values, ytrain.values, cat_features=cat_feature_ids)\n", "val_pool = Pool(Xval, yval, cat_features=cat_feature_ids) \n", "all_pool = Pool(X, y, cat_features=cat_feature_ids) \n", "\n", "# Spécification des paramètres d'entraînement du modèle\n", "model = CatBoostRegressor(iterations=300, \n", " depth=3, \n", " learning_rate=0.2, \n", " loss_function='RMSE')\n", "#Entraînement du modèle\n", "model.fit(train_pool, silent=True)" ], "execution_count": 412, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "" ] }, "metadata": { "tags": [] }, "execution_count": 412 } ] }, { "cell_type": "code", "metadata": { "id": "Ac894-uU00ou", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "ae7d6f56-cdd8-4434-d564-104defe40ef6" }, "source": [ "# Réalisation de la prédiction en utilisant le modèle obtenu\n", "trainpreds = model.predict(train_pool)\n", "valpreds = model.predict(val_pool)\n", "\n", "MAE(trainpreds,ytrain.values), MAE(valpreds,yval.values)" ], "execution_count": 413, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(10597.91476970661, 16062.060530924047)" ] }, "metadata": { "tags": [] }, "execution_count": 413 } ] }, { "cell_type": "code", "metadata": { "id": "8pS2SzLIXCmj", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "2a9c3e89-ee53-4595-9a23-6a07d704b520" }, "source": [ "# Spécification des paramètres d'entraînement du modèle\n", "model = CatBoostRegressor(iterations=300, \n", " depth=3, \n", " learning_rate=0.2, \n", " loss_function='RMSE')\n", "\n", "#Entraînement du modèle, avec un critère d'arrêt lorsque les performances en validation baissent: éviter le surentraînement (overfitting)\n", "model.fit(train_pool,eval_set = val_pool, early_stopping_rounds = 100, silent = True)" ], "execution_count": 414, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "" ] }, "metadata": { "tags": [] }, "execution_count": 414 } ] }, { "cell_type": "code", "metadata": { "id": "qOJRK7daRgy0", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "b3f61829-da8b-43c5-d54b-410f1e745b8d" }, "source": [ "# Réalisation de la prédiction en utilisant le modèle obtenu\n", "trainpreds = model.predict(train_pool)\n", "valpreds_catboost = model.predict(val_pool)\n", "\n", "#Pour éviter de prédire des valeurs anormales, on limite les prédictions au range du set d'entraînement\n", "trainpreds = trainpreds.clip(ytrain.min(), ytrain.max())\n", "valpreds = valpreds.clip(ytrain.min(), ytrain.max())\n", "\n", "MAE(trainpreds,ytrain.values), MAE(valpreds_catboost,yval.values)" ], "execution_count": 415, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(10597.91476970661, 16062.060530924047)" ] }, "metadata": { "tags": [] }, "execution_count": 415 } ] }, { "cell_type": "code", "metadata": { "id": "-4WuCGsNmTcD", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 456 }, "outputId": "cfbfefa9-dff4-4e71-9532-93c9bc4d8adc" }, "source": [ "display_results(yval.values, valpreds_lin, \"Régression linéaire\")\n", "display_results(yval.values, valpreds_poly, \"Régression polynomiale\")\n", "display_results(yval.values, valpreds_catboost, \"CatBoost\")" ], "execution_count": 416, "outputs": [ { "output_type": "stream", "text": [ "Mean average error: 21040.7804916666\n", "Mean average error: 19434.25335313196\n", "Mean average error: 16062.060530924047\n" ], "name": "stdout" }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAGDCAYAAAC8371AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEgQTDgAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "IryInRS1ZZna", "colab_type": "text" }, "source": [ "### Soumission des résultats" ] }, { "cell_type": "code", "metadata": { "id": "VlG-EhJ8ZM77", "colab_type": "code", "colab": {} }, "source": [ "Xtest = testset[newfeaturelist]\n", "test_pool = Pool(Xtest.values, cat_features=cat_feature_ids)\n", "\n", "#Pour la soumission, on peut éventuellement réentraîner le modèle sur toutes les données d'entraînement et de validation (pour espérer avoir un modèle plus générique. Cependant: attention à l'overfitting sans critère d'arrêt)\n", "model.fit(all_pool, silent = True)\n", "\n", "#Prédictions\n", "testpreds_catboost = model.predict(test_pool).clip(ytrain.min(), ytrain.max())\n", "\n", "submission = ids.to_frame()\n", "submission[\"SalePrice\"] = testpreds_catboost\n", "submission.to_csv(\"mysubmission_catboost.csv\", index = False)" ], "execution_count": 0, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "SoJxO1wcUh4p", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 287 }, "outputId": "4b64f5ed-1438-465e-8b3d-aba44df339ae" }, "source": [ "#Pour s'assurer de la plausibilité des résultats, on peut comparer leur distribution avec ceux du dataset d'entraînement\n", "plt.subplot(1,2,1)\n", "submission.SalePrice.hist()\n", "plt.subplot(1,2,2)\n", "ames.SalePrice.hist()" ], "execution_count": 418, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "" ] }, "metadata": { "tags": [] }, "execution_count": 418 }, { "output_type": "display_data", "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [] } } ] }, { "cell_type": "markdown", "metadata": { "id": "lMqBzDv9gMzb", "colab_type": "text" }, "source": [ "# Conclusion\n", "\n", "Les résultats avec le fichier de baseline (sample_submission.csv) ont été obtenus avec une régression linéaire sur l'année et le mosi de vente, la surface du lot et le nombre de chambres. Les résultats de la soumission sur Kaggle donnent: \n", "\n", "`MAE = 59346`\n", "\n", "Les trois soumissions proposées donnent:\n", "\n", "* régression linéaire: 20391\n", "* régression polynomiale: 19669\n", "* catboost: 14465\n", "\n", "Les algotihmes proposés prédisent donc bien mieux mieux les prix des maisons que la baseline.\n", "\n", "Le résultat obtenu avec catboost amène à la place 544/13783 au classement de la compétition Kaggle. Il reste donc de la marge de manoeuvre pour améliorer le modèle. (Le meilleur score publique est actuellement à 11824).\n", "\n", "Il existe de très nombreuses possibilités pour améliorer le modèle, dont par exemple:\n", "\n", "* Vérifier plus profondément les données et supprimer d'éventuelles anomalies\n", "* Tester d'autres ensembles de variables (en explorant plus profondément leurs relations et leur influence sur le prix)\n", "* Créer de nouvelles variables pertinentes. Par exemple, à partir de la variable catégorielle \"neighboorhood\" et du prix moyen de chaque quartier, il est possible d'extraire une variable numérique indiquant le \"standing\" du quartier. Les nombreuses variables indiquant une échelle de qualité (\"Excellent\",\"Typical\",\"Fair\", ...) peuvent être traduites en variables numériques. Des variables \"dummy\" peuvent être extraites sur d'autres catégories. Il pourrait également être intéressant de considérer [année de construction du garage] - [année de construction de la maison], par exemple pour détecter les garages \"intégrés\" (plus jolis que les ajouts postérieurs?) ou, au contraire, les travaux de re-valuation récents. Des variables simplifiées peuvent être utilisées: indiquer simplement la présence ou non d'une piscine ?\n", "* Les variables sont peut-être redondantes (elles apportent la même information), ce qui complexifie inutilement le modèle. Il peut être intéressant d'extraire un ensemble de variables apportant un maximum d'informations non-redondantes. E.g.: GarageArea et GarageCars apportent presque la même information.\n", "* Optimiser les hyperparamètres des modèles utilisés grâce à une k-fold cross-validation (régularisation de la régression linéaire, paramètres d'apprentissage d'autres algorithmes; e.g. nombre et profondeur des arbres pour CatBoost, nombre d'itération et taux d'apprentissage)\n", "* Tester d'autres algorithmes de machine learning (e.g.: support vector regression).\n", "* Entraîner plusieurs modèles, et calculer une moyenne pondérée des prédictions (on parle de model blending, ou de ensemble learning)\n", "\n", "Kaggle est une excellente source d'inspiration pour s'améliorer. On y trouve en effet de très nombreux notebooks (généralement en Python) proposant des idées intéressantes de feature engineering et de développement de modèles prédictifs. Par exemple, le notebook suivant est classé 26ème:\n", "https://www.kaggle.com/itslek/stack-blend-lrs-xgb-lgb-house-prices-k-v17\n", "La méthode propose certaines nouvelles variables prédictives simplifiées, ainsi que la moyenne pondérée de différents modèles prédictifs.\n", "\n" ] }, { "cell_type": "markdown", "metadata": { "id": "6zGX_8fbgPar", "colab_type": "text" }, "source": [ "## Bonus - Un exemple basique de model blending" ] }, { "cell_type": "code", "metadata": { "id": "q_C7oDL4XJbm", "colab_type": "code", "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "outputId": "c86bd5fd-e1c3-438e-9ec7-b25daf741585" }, "source": [ "w1, w2, w3 = 0.02, 0.08, 0.9\n", "valpreds_blend = w1*valpreds_lin + w2*valpreds_poly + w3*valpreds_catboost\n", "MAE(yval, valpreds_blend)" ], "execution_count": 419, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "15742.155442434165" ] }, "metadata": { "tags": [] }, "execution_count": 419 } ] }, { "cell_type": "code", "metadata": { "id": "aXQrS9F3dLx7", "colab_type": "code", "colab": {} }, "source": [ "testpreds_blend = w1*testpreds_lin + w2*testpreds_poly + w3*testpreds_catboost\n", "submission = ids.to_frame()\n", "submission[\"SalePrice\"] = testpreds_blend\n", "submission.to_csv(\"mysubmission_blend.csv\", index = False)\n" ], "execution_count": 0, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "UTUt50aQatG6", "colab_type": "text" }, "source": [ "* model blending: `MAE = 14273`\n", "Nouveau classement: 516/13783" ] } ] }