{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", " \n", "## [mlcourse.ai](https://mlcourse.ai) – Open Machine Learning Course \n", "\n", "Author: [Yury Kashnitskiy](https://yorko.github.io). Translated by [Sergey Oreshkov](https://www.linkedin.com/in/sergeoreshkov/). This material is subject to the terms and conditions of the [Creative Commons CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) license. Free use is permitted for any non-commercial purpose." ] }, { "cell_type": "markdown", "metadata": { "collapsed": true }, "source": [ "#
Assignment #8 (demo)\n", "\n", "##
Implementation of online regressor\n", " \n", "**Same assignment as a [Kaggle Kernel](https://www.kaggle.com/kashnitsky/a8-demo-implementing-online-regressor) + [solution](https://www.kaggle.com/kashnitsky/a8-demo-implementing-online-regressor-solution).**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Here we'll implement a regressor trained with stochastic gradient descent (SGD). Fill in the missing code. If you do evething right, you'll pass a simple embedded test." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##
Linear regression and Stochastic Gradient Descent" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd\n", "from sklearn.base import BaseEstimator\n", "from sklearn.metrics import log_loss, mean_squared_error, roc_auc_score\n", "from sklearn.model_selection import train_test_split\n", "from tqdm import tqdm\n", "\n", "%matplotlib inline\n", "import seaborn as sns\n", "from matplotlib import pyplot as plt\n", "from sklearn.preprocessing import StandardScaler" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Implement class `SGDRegressor`. Specification:\n", "- class is inherited from `sklearn.base.BaseEstimator`\n", "- constructor takes parameters `eta` – gradient step ($10^{-3}$ by default) and `n_epochs` – dataset pass count (3 by default)\n", "- constructor also creates `mse_` and `weights_` lists in order to track mean squared error and weight vector during gradient descent iterations\n", "- Class has `fit` and `predict` methods\n", "- The `fit` method takes matrix `X` and vector `y` (`numpy.array` objects) as parameters, appends column of ones to `X` on the left side, initializes weight vector `w` with **zeros** and then makes `n_epochs` iterations of weight updates (you may refer to this [article](https://medium.com/open-machine-learning-course/open-machine-learning-course-topic-8-vowpal-wabbit-fast-learning-with-gigabytes-of-data-60f750086237) for details), and for every iteration logs mean squared error and weight vector `w` in corresponding lists we created in the constructor. \n", "- Additionally the `fit` method will create `w_` variable to store weights which produce minimal mean squared error\n", "- The `fit` method returns current instance of the `SGDRegressor` class, i.e. `self`\n", "- The `predict` method takes `X` matrix, adds column of ones to the left side and returns prediction vector, using weight vector `w_`, created by the `fit` method." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "class SGDRegressor(BaseEstimator):\n", " # you code here\n", " def __init__(self):\n", " pass\n", "\n", " def fit(self, X, y):\n", " pass\n", "\n", " def predict(self, X):\n", " pass" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let's test out the algorithm on height/weight data. We will predict heights (in inches) based on weights (in lbs)." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "data_demo = pd.read_csv(\"../../data/weights_heights.csv\")" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "plt.scatter(data_demo[\"Weight\"], data_demo[\"Height\"])\n", "plt.xlabel(\"Weight (lbs)\")\n", "plt.ylabel(\"Height (Inch)\")\n", "plt.grid();" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X, y = data_demo[\"Weight\"].values, data_demo[\"Height\"].values" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Perform train/test split and scale data." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X_train, X_valid, y_train, y_valid = train_test_split(\n", " X, y, test_size=0.3, random_state=17\n", ")" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "scaler = StandardScaler()\n", "X_train_scaled = scaler.fit_transform(X_train.reshape([-1, 1]))\n", "X_valid_scaled = scaler.transform(X_valid.reshape([-1, 1]))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Train created `SGDRegressor` with `(X_train_scaled, y_train)` data. Leave default parameter values for now." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Draw a chart with training process – dependency of mean squared error from the i-th SGD iteration number." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Print the minimal value of mean squared error and the best weights vector." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Draw chart of model weights ($w_0$ and $w_1$) behavior during training." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Make a prediction for hold-out set `(X_valid_scaled, y_valid)` and check MSE value." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here\n", "sgd_holdout_mse = 10" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Do the same thing for `LinearRegression` class from `sklearn.linear_model`. Evaluate MSE for hold-out set." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# you code here\n", "linreg_holdout_mse = 9" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "try:\n", " assert (sgd_holdout_mse - linreg_holdout_mse) < 1e-4\n", " print(\"Correct!\")\n", "except AssertionError:\n", " print(\n", " \"Something's not good.\\n Linreg's holdout MSE: {}\"\n", " \"\\n SGD's holdout MSE: {}\".format(linreg_holdout_mse, sgd_holdout_mse)\n", " )" ] } ], "metadata": { "anaconda-cloud": {}, "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.7.1" } }, "nbformat": 4, "nbformat_minor": 2 }