{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Lesson 11 - Dimensions\n", "\n", "> Introduction to dimensionality reduction and visualisation." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/lewtun/dslectures/master?urlpath=lab/tree/notebooks%2Flesson11_visualisation.ipynb)[![slides](https://img.shields.io/static/v1?label=slides&message=2021-lesson11.pdf&color=blue&logo=Google-drive)](https://drive.google.com/open?id=1j_JaTfUbjoFUmtgkD1d_w0rQo61qSpbX)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Learning objectives\n", "In this lecture we will have a look at dimensionality reduction and data visualisation. When we look at the feature vector of a dataset it can contain many features and each feature corresponds to its own dimension. In the first part of the notebook we will explore a method to reduce the dimensionality of a dataset. In the second part we will look at a an algorithm to extract information from high-dimensional data and display its structure in two dimensions. Both these methods fall into the cateogory of unsupersvised algorithms. The learning objectives are to be able to answer the following questions:\n", "\n", "* What is dimensionality reduction?\n", "* When is dimensionality reduction useful?\n", "\n", "## References\n", "\n", "* Chapter 8: Dimensionality Reduction, Section PCA of _Hands-On Machine Learning with Scikit-Learn and TensorFlow_ by Aurèlien Geron\n", "\n", "## Homework\n", "As homework read the references, work carefully through the notebook and solve the exercises. This lecture covers several complex topics and it is important that you familiarise yourself by experimenting with the notebook." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Principal component analysis\n", "\n", "First we have a look at a method to reduce the dimensions of a dataset. There are two main reasons why we would like to reduce the dimensionality of such datasets:\n", "\n", "1. Some machine learning algorithms struggle with high dimensional data.\n", "2. It is hard to visualize high dimensional data. In practice it is hard to visualise datasets that have more than two or three dimensions.\n", "\n", "There is one very common approach to reduce the dimensionality of data: principal component analysis (PCA). The idea is that not all axis contain the same amount of variance and that there are even combination of axis that contain most variance. PCA seeks to find new coordinate axis such that the variance along the axis is maximised and ordered (the first principal component conatains most variance).\n", "\n", "
\n", "\n", "

Figure: PCA can look intimidating at the beginning.

\n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Imports" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# dalineplotrangling\n", "import pandas as pd\n", "import numpy as np\n", "from dslectures.core import get_dataset, convert_strings_to_categories, rmse, fill_missing_values_with_median\n", "from pathlib import Path\n", "from tqdm import tqdm\n", "\n", "# data viz\n", "import matplotlib.pyplot as plt\n", "import matplotlib as mpl\n", "import matplotlib.image as mpimg\n", "import seaborn as sns\n", "\n", "sns.set(color_codes=True)\n", "sns.set_palette(sns.color_palette(\"muted\"))\n", "\n", "# ml magic\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.ensemble import RandomForestRegressor\n", "from sklearn.metrics import r2_score\n", "from sklearn.preprocessing import StandardScaler\n", "from sklearn.decomposition import PCA\n", "from sklearn.model_selection import cross_validate\n", "\n", "#dslecture\n", "from dslectures.core import rmse" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# uncomment if running on Binder / your laptop\n", "# !pip install --upgrade dslectures" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load the data\n", "We will use the processed housing dataset for the principal component analysis. First we have to load it." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Dataset already exists at '../data/housing_processed.csv' and is not downloaded again.\n" ] } ], "source": [ "get_dataset('housing_processed.csv')" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "data_path = Path('../data/')" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "housing_data = pd.read_csv(data_path/'housing_processed.csv')" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "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", "
longitudelatitudehousing_median_agetotal_roomstotal_bedroomspopulationhouseholdsmedian_incomemedian_house_valuecitypostal_coderooms_per_householdbedrooms_per_householdbedrooms_per_roompopulation_per_householdocean_proximity_INLANDocean_proximity_<1H OCEANocean_proximity_NEAR BAYocean_proximity_NEAR OCEANocean_proximity_ISLAND
0-122.2337.8841.0880.0129.0322.0126.08.3252452600.069947056.9841271.0238100.1465912.55555600100
1-122.2237.8621.07099.01106.02401.01138.08.3014358500.0620946116.2381370.9718800.1557972.10984200100
2-122.2437.8552.01467.0190.0496.0177.07.2574352100.0620946188.2881361.0734460.1295162.80226000100
3-122.2537.8552.01274.0235.0558.0219.05.6431341300.0620946185.8173521.0730590.1844582.54794500100
4-122.2537.8552.01627.0280.0565.0259.03.8462342200.0620946186.2818531.0810810.1720962.18146700100
\n", "
" ], "text/plain": [ " longitude latitude housing_median_age total_rooms total_bedrooms \\\n", "0 -122.23 37.88 41.0 880.0 129.0 \n", "1 -122.22 37.86 21.0 7099.0 1106.0 \n", "2 -122.24 37.85 52.0 1467.0 190.0 \n", "3 -122.25 37.85 52.0 1274.0 235.0 \n", "4 -122.25 37.85 52.0 1627.0 280.0 \n", "\n", " population households median_income median_house_value city \\\n", "0 322.0 126.0 8.3252 452600.0 69 \n", "1 2401.0 1138.0 8.3014 358500.0 620 \n", "2 496.0 177.0 7.2574 352100.0 620 \n", "3 558.0 219.0 5.6431 341300.0 620 \n", "4 565.0 259.0 3.8462 342200.0 620 \n", "\n", " postal_code rooms_per_household bedrooms_per_household \\\n", "0 94705 6.984127 1.023810 \n", "1 94611 6.238137 0.971880 \n", "2 94618 8.288136 1.073446 \n", "3 94618 5.817352 1.073059 \n", "4 94618 6.281853 1.081081 \n", "\n", " bedrooms_per_room population_per_household ocean_proximity_INLAND \\\n", "0 0.146591 2.555556 0 \n", "1 0.155797 2.109842 0 \n", "2 0.129516 2.802260 0 \n", "3 0.184458 2.547945 0 \n", "4 0.172096 2.181467 0 \n", "\n", " ocean_proximity_<1H OCEAN ocean_proximity_NEAR BAY \\\n", "0 0 1 \n", "1 0 1 \n", "2 0 1 \n", "3 0 1 \n", "4 0 1 \n", "\n", " ocean_proximity_NEAR OCEAN ocean_proximity_ISLAND \n", "0 0 0 \n", "1 0 0 \n", "2 0 0 \n", "3 0 0 \n", "4 0 0 " ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "housing_data.head()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(19443, 20)" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "housing_data.shape" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "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", "
longitudelatitudehousing_median_agetotal_roomstotal_bedroomspopulationhouseholdsmedian_incomemedian_house_valuecitypostal_coderooms_per_householdbedrooms_per_householdbedrooms_per_roompopulation_per_householdocean_proximity_INLANDocean_proximity_<1H OCEANocean_proximity_NEAR BAYocean_proximity_NEAR OCEANocean_proximity_ISLAND
count19443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.00000019443.000000
mean-119.56036335.64673928.4351182617.678548538.1369641442.129970501.4273523.675099191793.406162541.62922493030.1456055.3402451.0917410.2148123.0959530.3314820.4399530.1067740.1215350.000257
std2.0026972.14533512.5045842179.553070420.1685321140.254218383.0642221.56968796775.724042260.7045121853.6843522.1904050.4297280.05666710.6790360.4707580.4963940.3088330.3267560.016035
min-124.35000032.5500001.0000002.0000002.0000003.0000002.0000000.49990014999.0000001.00000085344.0000000.8461540.3333330.1000000.7500000.0000000.0000000.0000000.0000000.000000
25%-121.76000033.93000018.0000001438.500000299.000000799.000000282.0000002.526500116700.000000328.00000091706.0000004.4123781.0061400.1779062.4496920.0000000.0000000.0000000.0000000.000000
50%-118.49000034.26000029.0000002111.000000436.0000001181.000000411.0000003.446400173400.000000545.00000092860.0000005.1804511.0482760.2045452.8411550.0000000.0000000.0000000.0000000.000000
75%-117.99000037.73000037.0000003119.000000644.0000001746.500000606.0000004.579750247100.000000770.00000094606.0000005.9637961.0977010.2404143.3082081.0000001.0000000.0000000.0000000.000000
max-114.31000041.95000052.00000039320.0000006445.00000035682.0000006082.00000015.000100499100.000000977.00000096161.000000132.53333334.0666671.0000001243.3333331.0000001.0000001.0000001.0000001.000000
\n", "
" ], "text/plain": [ " longitude latitude housing_median_age total_rooms \\\n", "count 19443.000000 19443.000000 19443.000000 19443.000000 \n", "mean -119.560363 35.646739 28.435118 2617.678548 \n", "std 2.002697 2.145335 12.504584 2179.553070 \n", "min -124.350000 32.550000 1.000000 2.000000 \n", "25% -121.760000 33.930000 18.000000 1438.500000 \n", "50% -118.490000 34.260000 29.000000 2111.000000 \n", "75% -117.990000 37.730000 37.000000 3119.000000 \n", "max -114.310000 41.950000 52.000000 39320.000000 \n", "\n", " total_bedrooms population households median_income \\\n", "count 19443.000000 19443.000000 19443.000000 19443.000000 \n", "mean 538.136964 1442.129970 501.427352 3.675099 \n", "std 420.168532 1140.254218 383.064222 1.569687 \n", "min 2.000000 3.000000 2.000000 0.499900 \n", "25% 299.000000 799.000000 282.000000 2.526500 \n", "50% 436.000000 1181.000000 411.000000 3.446400 \n", "75% 644.000000 1746.500000 606.000000 4.579750 \n", "max 6445.000000 35682.000000 6082.000000 15.000100 \n", "\n", " median_house_value city postal_code rooms_per_household \\\n", "count 19443.000000 19443.000000 19443.000000 19443.000000 \n", "mean 191793.406162 541.629224 93030.145605 5.340245 \n", "std 96775.724042 260.704512 1853.684352 2.190405 \n", "min 14999.000000 1.000000 85344.000000 0.846154 \n", "25% 116700.000000 328.000000 91706.000000 4.412378 \n", "50% 173400.000000 545.000000 92860.000000 5.180451 \n", "75% 247100.000000 770.000000 94606.000000 5.963796 \n", "max 499100.000000 977.000000 96161.000000 132.533333 \n", "\n", " bedrooms_per_household bedrooms_per_room population_per_household \\\n", "count 19443.000000 19443.000000 19443.000000 \n", "mean 1.091741 0.214812 3.095953 \n", "std 0.429728 0.056667 10.679036 \n", "min 0.333333 0.100000 0.750000 \n", "25% 1.006140 0.177906 2.449692 \n", "50% 1.048276 0.204545 2.841155 \n", "75% 1.097701 0.240414 3.308208 \n", "max 34.066667 1.000000 1243.333333 \n", "\n", " ocean_proximity_INLAND ocean_proximity_<1H OCEAN \\\n", "count 19443.000000 19443.000000 \n", "mean 0.331482 0.439953 \n", "std 0.470758 0.496394 \n", "min 0.000000 0.000000 \n", "25% 0.000000 0.000000 \n", "50% 0.000000 0.000000 \n", "75% 1.000000 1.000000 \n", "max 1.000000 1.000000 \n", "\n", " ocean_proximity_NEAR BAY ocean_proximity_NEAR OCEAN \\\n", "count 19443.000000 19443.000000 \n", "mean 0.106774 0.121535 \n", "std 0.308833 0.326756 \n", "min 0.000000 0.000000 \n", "25% 0.000000 0.000000 \n", "50% 0.000000 0.000000 \n", "75% 0.000000 0.000000 \n", "max 1.000000 1.000000 \n", "\n", " ocean_proximity_ISLAND \n", "count 19443.000000 \n", "mean 0.000257 \n", "std 0.016035 \n", "min 0.000000 \n", "25% 0.000000 \n", "50% 0.000000 \n", "75% 0.000000 \n", "max 1.000000 " ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "housing_data.describe()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We drop the two columns that contain categorical data with many categories. We could create extra columns for each categoriy but since there are many of them that would create a lot of extra features." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "housing_data.drop(['city', 'postal_code'], axis=1, inplace=True)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Then we split the data into features and labels as usual:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X = housing_data.drop('median_house_value', axis=1)\n", "y = housing_data['median_house_value']\n", "feature_labels = X.columns" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Baseline\n", "First we train the a Random Forest with the settings we tuned in lesson 6 as a baseline for future model." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "rf = RandomForestRegressor(n_estimators=100, max_features='sqrt', n_jobs=-1, random_state=42)\n", " \n", "results = cross_validate(rf, X, y,\n", " cv=5,\n", " return_train_score=True,\n", " scoring='neg_root_mean_squared_error')\n", "rmse_full = -np.mean(results[\"test_score\"])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We get an RMSE of roughly 60'000:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "RMSE: 60301.958010516806\n" ] } ], "source": [ "print(f'RMSE: {rmse_full}')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Standardisation\n", "PCA works by finding new axes in the dataset that cover the most variance. Since variance is the average squared distance of each sample to the sample mean this depends on the scale of the feature. To illustrate this think of two columns that contain the height of the person measured in centimeter and meter. Although height described in these two columns are the same the variance will be 10'000 times larger in the centimeter column (100^2) and the standard deviation 100 times larger." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "rand_values = np.random.randn(1000)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "1.0254503528227774" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "rand_values.var()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "10254.503528227773" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "(100*rand_values).var()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For this reason it is important to scale the features such that they are comparable to each other. A common approach to do this is to transform the data such that the mean is zero and the standard deviation is one. The following formula achieves this:\n", "\n", "$x_{std}=\\frac{x_{old}-\\mu}{\\sigma}$\n", "\n", "Scikit-learn provides a function to do this for us called `StandardScaler`. We can do this in one line with the function `.fit_transform()`. Similar to the function `fit_predict()` it combines the process of fitting which means calculating the mean and standard deviation of the dataset and then transforming the data by applying the above described formula." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X_std = StandardScaler().fit_transform(X)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(19443, 17)" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_std.shape" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the mean is zero (or close enough) and the standard deviation is one:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([-4.44386137e-16, 2.19854194e-15, 9.35549763e-17, -2.48505406e-17,\n", " -1.02325755e-16, 7.74752147e-17, -1.24252703e-17, -1.46179650e-16,\n", " -9.94021623e-17, 1.57874022e-16, 7.01662322e-17, 2.92359301e-17,\n", " 1.60797615e-16, 1.43256057e-16, 6.13954532e-17, -3.14286248e-17,\n", " 1.16943720e-17])" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_std.mean(axis=0)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_std.std(axis=0)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Principal component analysis" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we apply principal component analysis to the standardised features. After initialising a `PCA` object we can use the `.fit_transform()` to find the principal components and then transform the dataset into these new coordinates.\n", "\n", "> Note: You should never `fit_transform` the test set. Use it on the training set and apply the mean and standard deviation with `transform` to the test set. This is true for all data preprocessing." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "pca = PCA()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X_pca = pca.fit_transform(X_std)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the transformed dataset still has the same shape. However the columns are no longer the features but the coordinates in the new principal component coordinates." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(19443, 17)" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_pca.shape" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Explained variance\n", "The explained variance is an important concept in principal component analysis. It tells us how much variance of the whole dataset is contained along one principal component axis. The component with the more variation can encode more information than features with little variation. The explained variance is stored in the `pca` object and we can use it to visualise the distribution. The `explained_variance_ratio_` tells us the percentage of variance each component contains from the total variance." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([2.32152692e-01, 1.58210508e-01, 1.27146991e-01, 1.00792055e-01,\n", " 7.93230339e-02, 7.26018573e-02, 5.93200071e-02, 5.87781344e-02,\n", " 4.47812090e-02, 3.66692458e-02, 1.63533745e-02, 7.52994128e-03,\n", " 2.65596682e-03, 1.50076877e-03, 1.26555424e-03, 9.18660939e-04,\n", " 1.01541990e-32])" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pca.explained_variance_ratio_" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "pca_labels = [f'principal component {i+1}' for i in range(len(pca.explained_variance_ratio_))]" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAFxCAYAAABweRMUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO3de3hM5/428DtyEBE/cUhCtT9KEURKRcWx4pADOTjE4VXsXapFsUuroahT7dqqolTpbl/tLkqESOkhbKVaRIu2km6hqs40iSBCEonM8/7Rd2aLpJKZPJHvWPfnulzNrJl15540+Was9VjjoJRSICKiB16Vyi5ARET3Bwc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgkwg9evRASkrKPR/z9ttvIyEhQdvnfP755xEfH68t706RkZG4fv16hWTrlJ2djZEjR1pu20tvso1TZRcgKqu//e1vlV2hzD799NPKrlAmWVlZRX7R2ktvsg0HPtls165dWLlyJQoKCuDq6oro6Gi0bdsW06dPR05ODt5++22cOHECI0eOxJo1a/Dll1/ixIkTuHz5MjIzM+Hj44MFCxbA3d3dkmkymfD3v/8dR44cwc2bN6GUwuuvv4527dph2rRpaNq0KUaPHo3WrVvjueeew759+5Ceno6RI0fir3/9KwAgLi4O69evh8lkgoeHB2bNmoUmTZogLS0N06ZNQ3p6Oh566CFkZmYWe07Z2dl46qmnsH37dnh6egIABg8ejBdeeAH/+7//i3nz5iEnJwfp6enw8fHB0qVLUbVqVfj6+qJnz544duwYFi9ejKioKCQlJcHV1RVz5szB6dOnkZWVherVq2Px4sVo3LgxRowYgTZt2uCHH37ApUuX0K5dO/zjH/9AlSpVsHv3bixduhQmkwlubm6YO3cufHx88MMPP2Dx4sXIzc2Fg4MDJk6ciMDAwGLP4+4+x48fR2xsLAoKCpCVlYUxY8Zg2LBhmD59OvLy8hAZGYn4+Hi0bNkSSUlJqF27NlasWIHPP/8cjo6OePTRRzFr1izL14TslCKywalTp1RYWJi6cuWKUkqpX375RXXu3FndvHlT3bx5UwUFBan4+HjVt29ftXXrVqWUUsuWLVPdunVTGRkZqrCwUE2ZMkUtXLhQKaVUYGCgSk5OVj/88IOaOHGiKiwsVEop9d5776nnn39eKaVUdHS0+uCDD5RSSjVr1kytWbNGKaVUSkqK8vX1VXl5eeq7775Tw4YNUzk5OUoppb799lsVGhqqlFJq/PjxKiYmRiml1OnTp1WbNm3U5s2biz23V155xfJ5fv31V9W9e3dVWFioFi5cqBISEpRSSuXn56uwsDCVmJho6bNlyxZLRrNmzVRmZqb68ssv1fz58y3bZ82apebNm6eUUmr48OFq0qRJqrCwUGVnZ6suXbqopKQklZGRodq1a6eOHj2qlFJq+/btavTo0eratWsqKChInTt3Timl1O+//666deumLly4UOw53Nnnxo0bavDgwZb/Vz/++KNq06aNUkqpc+fOWT6+s/emTZvUkCFD1M2bNy3/70aNGlXCdwLZE77CJ5uYX1mbX1UDgIODA86ePQsfHx/ExMRg8ODBiIiIQHh4uOUxISEhqFu3LgAgKioKf//73xEdHW25v23btqhZsyY2bNiAc+fO4bvvvkP16tVL7NCzZ08AQKtWrZCfn4+cnBx8/fXXOHPmDIYOHWp5XFZWFq5du4b9+/dbPlfDhg3RoUOHEnMHDRqEuXPnYvTo0di8eTMGDBiAKlWqYOrUqdi3bx/ef/99nD59Gunp6cjJybHs5+/vXywrJCQEjzzyCNasWYMzZ87g+++/R9u2bS33BwYGokqVKnB3d0fDhg2RlZWFH374AU2bNkWLFi0AAEFBQQgKCsKePXuQkZGBF154ocjX/Pjx43jooYeKfW5zn+rVq2PVqlXYs2cPTp8+jWPHjhXpXZJvvvkGAwYMgJubGwBg5MiRWLVqFfLz8+Hi4nLPfUkuDnyyiclkQseOHbF06VLLtkuXLsHLywsAcOrUKXh4eCA1NbXIkHB0dCySUaVK0XUDX3/9NRYsWIBnnnkGPXv2ROPGjbF169YSO1StWhXAH0MPAJRSMJlMiIyMxNSpUy2fIz09HTVr1oSDgwPUHZeOcnIq+dvf398ft2/fRnJyMj777DNs2LABADBlyhQUFhYiNDQU3bt3x6VLl4rkmYfjnT755BNs3LgRTz/9NMLDw+Hh4YHz589b7nd1dbV8bO7n5ORkeU7m53X8+HEUFhaiSZMmiIuLs9yXlpaG2rVrl/g8zH1+//13DBkyBIMHD0a7du0QEhKC3bt3l7jPnZ/zTiaTCbdv377nPiQfV+mQTQICArBv3z6cPHkSALBnzx5ERETg1q1bOH/+PBYsWIDVq1ejcePGWLx4sWW/r776CtnZ2TCZTNi4cWOx48/79u1DYGAghg0bhtatW2Pnzp0oLCwsc6/OnTvj888/R3p6OgBg/fr1+Mtf/gIA6Nq1K2JjYwEAFy9exHffffenOYMGDcL8+fPRvHlzy6vnvXv34oUXXkCfPn3g4OCAI0eOlNpt79696N+/PwYNGoRHH30Uu3btKnWfxx9/HCdPnsSJEycA/PE1mzp1Ktq0aYMzZ87g4MGDAIDU1FQEBwdbnuuf+fnnn1G7dm2MHz8eXbt2tQz7wsJCODk5obCwsNiA79KlC+Lj4y1/E1izZg3at2/PV/d2jq/wySZNmzbFvHnzMGXKFMur0pUrV8LFxQUvvfQSRo8ejWbNmuG1115DeHg4OnXqBACoW7cuxowZg6tXr6J9+/YYO3ZskdyhQ4fi5ZdfRnh4OBwdHeHv748dO3bAZDKVqVfXrl0xZswYjBo1Cg4ODnB3d8c777wDBwcHzJ49G9OnT0doaCjq1asHHx+fP83p168flixZgiVLlli2TZ48GS+88AJq1qyJatWqoX379jh79uw9+4waNQqvvfYa4uPj4ejoiFatWuGXX3655z5169bF4sWLER0djcLCQri7uyMmJga1a9fGsmXLsGjRIty6dQtKKSxatAgNGjS4Z17nzp2xadMmhISEoFq1avDz80Pt2rVx5swZNGzYEC1btkRoaCjWr19v2ScqKgqXLl3CoEGDYDKZ0LBhwyK/uMk+Oai7f7UTVZDly5fj6tWreO211yq7CpEh8ZAOEZFB8BU+EZFB8BU+EZFBcOATERkEBz4RkUFw4BMRGYTodfhXr96EycRzykREZVGligNq1Sr5UiSA8IFvMikOfCIiTXhIh4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIESvwweA2jVd4ejibPP+hfkFuJKVp7EREZF9Ej/wHV2ckbFyrc37e44bDoADn4iIh3SIiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMokwDf9u2bejTpw969+6NdevWFbt/586diIyMREREBMaPH4+srCwAwMWLF/H0008jJCQE48aNw82bN/W2JyKiMit14KelpSEmJgaffPIJPv30U8TGxuLXX3+13H/jxg3MmTMH//znP7F161Y0b94cy5cvBwDMnTsXw4YNQ2JiInx9ffHuu+9W3DMhIqJ7KnXg79+/HwEBAfDw8ICbmxuCg4ORmJhoub+goABz5syBt7c3AKB58+a4dOkSCgoKcPDgQQQHBwMABgwYUGQ/IiK6v0od+Onp6fD09LTc9vLyQlpamuV2rVq10KtXLwBAXl4e/vnPf6JXr164evUq3N3d4eTkBADw9PQssh8REd1fTqU9QClVbJuDg0OxbdnZ2Rg/fjx8fHzQv3//Eod7SfvdS5067lY9/s94etbQkkNEZM9KHfje3t44dOiQ5XZ6ejq8vLyKPCY9PR2jR49GQEAAXn31VQBA7dq1cePGDRQWFsLR0REZGRnF9itNZuYNLUM/IyO73BlERNJVqeJwz5lZ6iGdTp06ISkpCVeuXEFubi527NiBbt26We4vLCzE2LFjERoaihkzZlhexTs7O8Pf3x9ffPEFACAhIaHIfkREdH+V6RX+5MmTMXLkSBQUFCAqKgp+fn4YM2YMJk2ahN9//x1Hjx5FYWEhtm/fDgDw9fXFggULMHv2bEybNg0rV65E/fr1sWTJkgp/QkREVDIHVdJBeiHMh3QyVq61OcNz3HAe0iEiQyj3IR0iInowcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERmEU2UXuN9q16wKRxcXm/cvzM/HlaxbGhsREd0fhhv4ji4uuPTuDJv3rz9+AQAOfCKyPzykQ0RkEBz4REQGUaaBv23bNvTp0we9e/fGunXr/vRx0dHRiI+Pt9xOSEhAly5dEBkZicjISMTExJS/MRER2aTUY/hpaWmIiYlBfHw8XFxcMHToUHTo0AGPPfZYkcfMnj0bSUlJ6NChg2V7SkoKpk2bhrCwsIppT0REZVbqK/z9+/cjICAAHh4ecHNzQ3BwMBITE4s8Ztu2bejZsydCQ0OLbE9JSUFCQgIiIiLw8ssvIysrS297IiIqs1Jf4aenp8PT09Ny28vLC8nJyUUe8+yzzwIADh8+XGS7p6cnnnvuOfj5+WHJkiWYN28e3nrrrTKXq1PHvcyPvRdPzxpacioqj4jofih14Culim1zcHAoU/iKFSssHz/77LPo1auXFdWAzMwbWoZ+Rka25WMdw/rOPCIiKapUcbjnzCz1kI63tzcuX75suZ2eng4vL69SP3F2djY++ugjy22lFJycDLfsn4hIjFIHfqdOnZCUlIQrV64gNzcXO3bsQLdu3UoNdnNzwwcffIAjR44AANauXYvevXuXvzEREdmk1Jfc3t7emDx5MkaOHImCggJERUXBz88PY8aMwaRJk9C6desS93N0dMTSpUsxZ84c5OXloVGjRli0aJH2J0BERGXjoEo6SC+E+Rh+xsq1Nmd4jhte7Bh+eS+twGP4RCRRuY/hExHRg4EDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMgu85WE61arrAyaWqTfvezr+Fq1n5mhsREZWMA7+cnFyq4sdV4Tbt23bsNgAc+ER0f/CQDhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEBz4RkUFw4BMRGUSZBv62bdvQp08f9O7dG+vWrfvTx0VHRyM+Pt5y++LFi3j66acREhKCcePG4ebNm+VvTERENil14KelpSEmJgaffPIJPv30U8TGxuLXX38t9pixY8ciMTGxyPa5c+di2LBhSExMhK+vL95991297YmIqMxKHfj79+9HQEAAPDw84ObmhuDg4GKDfdu2bejZsydCQ0Mt2woKCnDw4EEEBwcDAAYMGFBsPyIiun+cSntAeno6PD09Lbe9vLyQnJxc5DHPPvssAODw4cOWbVevXoW7uzucnP74FJ6enkhLS9NS+kHlUdMFzi5Vbd6/IP8WrmXla2xERA+SUge+UqrYNgcHh1KDbd3vTnXquFv1+D/j6VlDS05F5N2d9cX/7WNzVp/RX8DT0/ZfGET0YCt14Ht7e+PQoUOW2+np6fDy8io1uHbt2rhx4wYKCwvh6OiIjIyMMu13p8zMG1qGfkZGtuVjHcNaZ15FdiMiY6lSxeGeM7PUY/idOnVCUlISrly5gtzcXOzYsQPdunUr9RM7OzvD398fX3zxBQAgISGhTPsREVHFKHXge3t7Y/LkyRg5ciT69euHsLAw+Pn5YcyYMUhJSbnnvrNnz8bGjRvRp08fHDp0CC+++KK24kREZJ1SD+kAQHh4OMLDw4tse//994s9buHChUVuN2jQAGvWrClHPSIi0oX/0paIyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDIIDn4jIIDjwiYgMggOfiMggOPCJiAyCA5+IyCDKdLVMsk81PZzh4uxq8/75BXnIulagsRERVSYO/AeYi7MrVv8ryOb9R/1lBwAOfKIHBQ/pEBEZBAc+EZFBcOATERkEBz4RkUHwpC2VWXlW/XDFD1Hl48CnMnNxdsXrscE27TtzyHZwxQ9R5eIhHSIig+DAJyIyCA58IiKD4MAnIjIIDnwiIoPgwCciMggOfCIig+DAJyIyCA58IiKD4MAnIjIIXlqBKkUNDxe4Ole1ef+8glvIvpavsRHRg48DnyqFq3NVhH76f2ze/8vI9cjGfwd+DQ9XuDo725yXV1CA7Gt5Nu9PZA848OmB4OrsjD5b/mHz/l/0j0Y2OPDpwcZj+EREBsGBT0RkEGUa+Nu2bUOfPn3Qu3dvrFu3rtj9qampGDhwIIKDgzFjxgzcvn0bAJCQkIAuXbogMjISkZGRiImJ0dueiIjKrNRj+GlpaYiJiUF8fDxcXFwwdOhQdOjQAY899pjlMVOnTsXrr7+ONm3a4NVXX8XGjRsxbNgwpKSkYNq0aQgLC6vQJ0FERKUr9RX+/v37ERAQAA8PD7i5uSE4OBiJiYmW+y9cuIC8vDy0adMGADBgwADL/SkpKUhISEBERARefvllZGVlVdDTICKi0pT6Cj89PR2enp6W215eXkhOTv7T+z09PZGWlmb5+LnnnoOfnx+WLFmCefPm4a233ipzuTp13Mv82Hvx9KyhJaci8iR3050nuVtF5BFJU+rAV0oV2+bg4FCm+1esWGHZ9uyzz6JXr15WlcvMvKFl6GdkZFs+1vFDrTNPcjfdeZK73Z1HZI+qVHG458ws9ZCOt7c3Ll++bLmdnp4OLy+vP70/IyMDXl5eyM7OxkcffWTZrpSCkxOX/RMRVZZSB36nTp2QlJSEK1euIDc3Fzt27EC3bt0s9zdo0ABVq1bF4cOHAfyxMqdbt25wc3PDBx98gCNHjgAA1q5di969e1fQ0yAiotKU+pLb29sbkydPxsiRI1FQUICoqCj4+flhzJgxmDRpElq3bo3Fixdj5syZuHnzJlq2bImRI0fC0dERS5cuxZw5c5CXl4dGjRph0aJF9+M5ERFRCcp0jCU8PBzh4eFFtr3//vuWj318fLBp06Zi+/n7+2PLli3lrEh0/5Xn2jy8Lg9JxYPqRCVwdXZG2Ob/a9O+nw0czevykEi8tAIRkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcJUOUQWr4VENrs62/6jlFdxG9rVcjY3IqDjwiSqYq7MTwjdttnn/bVEDwav8kA48pENEZBAc+EREBsGBT0RkEBz4REQGwYFPRGQQHPhERAbBgU9EZBAc+EREBsGBT0RkEBz4REQGwYFPRGQQHPhERAbBgU9EZBAc+EREBsGBT0RkEBz4REQGwTdAIbIzfActshUHPpGdcXV2Qv/Nu23ef8vAQL6DlkHxkA4RkUFw4BMRGQQHPhGRQXDgExEZBAc+EZFBcOATERkEl2USGVwNDze4OjvavH9eQSGyr+VobEQVhQOfyOBcnR0xZPMvNu8fO7AZ1/XbCR7SISIyCA58IiKDKNPA37ZtG/r06YPevXtj3bp1xe5PTU3FwIEDERwcjBkzZuD27dsAgIsXL+Lpp59GSEgIxo0bh5s3b+ptT0REZVbqwE9LS0NMTAw++eQTfPrpp4iNjcWvv/5a5DFTp07FrFmzsH37diilsHHjRgDA3LlzMWzYMCQmJsLX1xfvvvtuxTwLIhKhpkd1eHrWsPlPTY/qlf0UHmilnrTdv38/AgIC4OHhAQAIDg5GYmIiJkyYAAC4cOEC8vLy0KZNGwDAgAEDsGzZMgwaNAgHDx7EihUrLNuHDx+OqVOnlrlclSoOf/y3Rvm+Ccw5Zo41PLTmudTw0pZVzd32rJLy3Kt7a82r6WZ73t1ZXtXq2pxVYp7b/2jOc9eY5WZzVkl5nm6umvPKt37DnOfiXAWfJF62OWdYSN0i3f7nf9zg7Gz7keeCAhOuX//vCqKaNdzg5GJ73u18E7Ky/5vnUaMaHF1sW+FUmF+Ia9l6r1p69//XuzkopdS9HvDee+8hJycHkydPBgDExcUhOTkZ8+fPBwD8+OOPWLRoEdavXw8AOHPmDJ577jmsWbMGUVFR+OabbwAAt2/fRps2bfDzzz+X+0kREZH1Sv1VV9LvAwcHh1LvL20/IiK6v0od+N7e3rh8+b9/RUtPT4eXl9ef3p+RkQEvLy/Url0bN27cQGFhYZHtRERUOUod+J06dUJSUhKuXLmC3Nxc7NixA926dbPc36BBA1StWhWHDx8GACQkJKBbt25wdnaGv78/vvjiiyLbiYiocpR6DB/4Y1nme++9h4KCAkRFRWHMmDEYM2YMJk2ahNatW+PYsWOYOXMmbt68iZYtW+KNN96Ai4sLLly4gGnTpiEzMxP169fHkiVLULNmzfvxvIiI6C5lGvhERGT/+C9tiYgMggOfiMggOPCJiAyCA5+IyCA48ImIDMJxzpw5cyq7RGVJTU1FSkoK6tati6pVq1q27969G48++qhVWTk5OVBKwdHREfv378fOnTtRpUoV1KtXT0vX1atXo23btjbtm5aWBnf3P64L89tvv+Gzzz7D0aNHUbduXdSoUcOmzJMnT8LV1RXOzs74/vvvkZiYiPz8fDz88MNW5Xz44Ydo1aoVHB1tf8elu507dw45OTmoUaMGEhMTsX79ely9ehU+Pj425R09ehQbNmzA559/jn379uHUqVPw8vKy6Wt38uRJbNiwAdu2bcOePXtw9OhRuLq6wtu7fNc80mXnzp3Yt28fatWqVWQJdWxsLHx9fa3KOn36NG7dugV3d3fExcVhy5YtuHHjBpo2baql68KFC9GlSxeb9k1OTrZ8zZOSkrBu3TocOHAA1atXt+ln9ttvv0WtWrVQtWpVJCQkYOPGjbh8+TJatmxpddbrr7+Oxx9/HK6u5bteUkkMuyzzX//6FzZu3IhHHnkEKSkpWLx4MTp27AgA6N+/P7Zs2VLmrM8++wzz58+Hi4sLhgwZgu3bt6Nr167Yu3cvRowYgUGDBlnV7Z133im2bcOGDRg6dCgAWC5cV1bm55OYmIiFCxeid+/eUEph9+7dmDp1KkJCQqzK++CDDxAXF4cPP/wQu3btwsaNG9G1a1ckJSUhODgYzz//fJmzHn/8cTRp0gRz585F69atrepRko8++ggff/wxACAoKAjJycno27cvvv32WzRt2hQvvfSSVXlxcXGIjY1Fhw4dsG/fPnTu3Bm3bt3CV199hRkzZqBXr15lzlq3bh02btyI4OBgeHp6AvjjX6Dv2LEDERERGDVqlFXdLl68eM/7H3roIavyFi9ejJ9//hlNmjTBl19+iejoaERGRgKw/mfio48+wpo1a2AymRAQEIBLly6hd+/e2LVrF5544gm88MILVnWbPn16sW27du1Cjx49AABvvPGGVXnm57Nu3Tps2LABAwcOBABs2bIFgwYNwvDhw8uctWDBAqSmpiImJgbr1q1DSkoKevbsiW+++QYPP/wwZs6caVU3f39/1KlTBy+99BKCgoKs2rdUyk507txZ+fj4FPvTvHlz5ePjY3VeWFiYysnJUUopdfjwYdW5c2d18OBBpZRSkZGRVmWFh4erK1euqNTUVOXr66uuXbumlFIqOztbRUREWN1t4sSJqkOHDiomJkYtX75cLV++XHXu3NnysbX69eunlFKqf//+6vz585btGRkZKjQ01Oq84OBgy9euX79+KisrSymlVG5urgoODrYqKzIyUu3fv18FBQWpSZMmqZ9++snqPnfq27evysnJUVevXlVPPPGEunHjhlJKqfz8fBUWFmZ1Xr9+/VReXp5SSqmsrCw1ZswYpZRSly5dsjovKCjI8nW7U05OjtVfN6X++B729fVVPXr0UIGBgUX+9OjRw6a8goICpZRSp06dUoGBgeqLL75QSln/MxEWFqby8vLU+fPnVZs2bSxfw1u3bqnw8HCruy1cuFAFBASojz76SMXHx6v4+HjVvXt3y8fWMv9MREREqCtXrli2Z2dnW/3/IjQ0VN2+fVsp9cfX6datW0oppW7fvq1CQkKs7hYZGalOnDihnn76aRUVFaU+//xzlZuba3VOSezmPW03b96MkSNHYsWKFXjssce0ZFarVg0A8MQTT2DJkiV48cUXsXr1aqsv8qaUQq1atVCrVi1ERkZa/ipcvXp1y7WErLFs2TJs2rQJCQkJmDt3Lpo0aYKdO3da/cr+bm5ubqhfv77ldp06dWy6oJ2bmxtMJhMAwN3dHS4uLgAAR0dHODlZ9y3l4OCAjh074rPPPkNCQgLmz5+Pq1evol27dqhXrx6mTJliVZ76/4fVqlSpYulk/q/5jXmskZeXZ3lOLi4uuHTpEgCgXr16JV4g8F6cnJxK7JCXlwdnZ2eru61fvx7Dhg3D7Nmz0a5dO6v3v5tSyvL90KhRI7z33nt45plnULt2bau/T0wmE1xcXNCgQQOMGjWqyCFTW34moqOj0a1bNyxduhRTpkxBhw4d8K9//Qv9+/e3Ogv44+q9JpMJderUgdsdl652cXGxfO+UlaurKzIzM+Hl5YU6deogJycHLi4uyM3NtfrnAfjjZ+Kxxx7D2rVrsX//fsTGxmLBggVo1KgR6tWrh7feesvqTDO7Gfje3t549dVXsWzZMixbtqzcee3atcPkyZg45ksAABU7SURBVJMxYcIENGnSBE8++SRee+01PPPMM1Z/Q7Zt2xYvvfQS3nzzTbz++usA/rhM9Jtvvgl/f3+b+kVFRaF9+/aYOXMmgoODbcowy87ORp8+fWAymfDGG29gxowZ+Pnnn/HOO+/YNCiGDh2KQYMGYdiwYfD398fYsWMRGBiInTt3IiIiwqos89B0dnbGoEGDMGjQIKSlpeGnn37CuXPnrO7Wv39/hIaG4vbt2xg8eDCef/55BAUF4euvv8ZTTz1ldV6XLl3w3HPPoWfPnvjqq68QGBiI9PR0vPbaa2jfvr1VWWPHjkW/fv3QsWPHIod0Dhw4YLn8uDXc3d3x+uuvIy4uTsvADwkJwYgRIzBt2jT4+fmhadOmePvttzFhwgTk5+dblRUUFIThw4fj448/xsSJEwHAcgmW0NBQm/p17NgRLVq0wOzZs/H111/b9IvDrFatWnjqqafg4OCA2bNnY+HChUhKSsKbb75p9SHOCRMmICoqCn379kXjxo0xYsQIdOzYEXv37sWzzz5rdbc7X0h06tQJnTp1QkFBAY4fP27Tz8SdDHsM32QyYfPmzWjevDn8/Pws25OTk7Fq1Sqr3p2rsLAQ27ZtQ79+/Szb/vOf/+DIkSMYOnSo1a8Y7s5etmwZduzYgS+//NLmnGvXriElJQW3bt1Cr1698PXXX+PChQsYOnSoTSdMDx06hMTERJw5cwaFhYWoW7cuAgMDrf5hXrFihdXHc0tz9uxZmEwmNGrUCN9++y12794NHx8fDBo0yKa/vW3atAmpqalo2bIlBg4ciKysLPz4448IDAy0ultaWhqSkpKQnp4OpRS8vb3RsWNHMSdtk5KS4OXlhSZNmli2Xbp0CatXr8aMGTOsyjp48GCRX4q//fYbzp07Z9Mv3rvFxcXhyy+/xOrVq8uV89tvv+H69eto06YNDh8+jOzsbHTv3t3qnHPnzmHnzp3Ffh7unC1lFRcXZ/V5v7Iy7MAnIjIarsMnIjIIuzmGT2TvunTpgszMzGLbzSdLU1NTH5g8yd1050nuVlKIXVm1alWxbW+99ZaIPMndpOdJ7qYr7/fff1dBQUHqxIkTNvewlzzJ3XTnSe52N7s5hr948WJkZmYW+ccWwB/Lq5KTk7F9+/ZKy5PcTXqe5G4Vkbdnzx5s3rxZy0oz6XmSu+nOk9ztTnYz8JOTk3Hy5EksW7YMkyZNsmx3dHSEn58fGjVqVGl5krtJz5PcrSLyiCqT3Qx8s+zsbJuv/1LReZK7Sc+T3K0i8ogqg92dtN25cycWLlyI69evAyj/iQydeZK7Sc+T3K0i8ogqhfazAhWsR48e6vjx4yLzJHeTnie5W0XkEVUGu1uH7+3tjWbNmonMk9xNep7kbrrz3nvvvWLblixZ8kDmSe6mO09yNzO7O4a/YMECpKWloXPnzkUuyHTnZQ0qK09yN+l5krvpypO+goirrx7M53onuzuGf+PGDVSvXh0//fRTke22/iDrzJPcTXqe5G668oKCgnDy5EkcOHAATz75pGW7o6OjTdcTkpwnuZvuPMnd7mZ3r/DNsrKyirwjj6Q8yd2k50nupitP+goirr6q/KyKyAPs8Fo6x44dQ0hICCIjI5GWlobevXvjP//5j4g8yd2k50nupjtv586d6NChA1q0aIEWLVrAx8cHLVq0sLmb5DzJ3XTnSe5mUbnnjK03bNgw9euvv1regWfv3r1q4MCBIvIkd5OeJ7mb7jzpK4i4+qrysyoiTyk7XKWTm5tb5FrdnTt3tvrNGSoqT3I36XmSu+nOk7yCSHee5G668yR3M7O7k7YeHh44duyY5Y0stm7dWq5jqjrzJHeTnie5m+68Vq1aYdKkSdpWEEnOk9xNd57kbmZ2d9L27NmziI6ORkpKClxdXdGwYUO8+eabaNy4caXnSe4mPU9yN91506dPL3H7G2+8YVM3yXmSu+nOk9zNzO4GvllOTg5MJhPc3d3F5UnuJj1PcjfdeRJXEFVUnuRuuvMkd7O7QzpHjx7FqlWrkJWVVeTNfj/++ONKz5PcTXqe5G66844dO4YXX3wReXl5iI2NxfDhw7F06VK0atXKpm6S8yR3050nuZuF1lPA90FYWJhas2aNOnDggPruu+8sfyTkSe4mPU9yN915klcQ6c6T3E13nuRuZnb3Ct/V1RXDhw8XmSe5m/Q8yd1055W04ucf//jHA5knuZvuPMndzOxu4Hfp0gVr1qxBly5dipy5fuihhyo9T3I36XmSu+nOk7yCSHee5G668yR3M7O7k7Z3XkzIzMHBAV999VWl50nuJj1PcjfdeZJXEOnOk9xNd57kbmZ2N/CJHhSSVxDpzpPcTXee5G52d0jnypUrmDdvHpKSklBYWIiAgADMmTMHdevWrfQ8yd2k50nupjtP8goi3XmSu+nOk9zNzO5e4U+YMAFt27bFkCFDYDKZEBsbi0OHDpX4ZgH3O09yN+l5krvpzgsPD8eQIUPQtGlTy/FZAEUuhfug5EnupjtPcjeLcq3xqQQRERHFtoWFhYnIk9xNep7kbrrzoqKibO5hb3mSu+nOk9zNzO4unubg4IBLly5Zbl+8eBFOTrYfmdKZJ7mb9DzJ3XTnmVf8nDp1ChcvXrT8sZXkPMnddOdJ7mZmd4d0du/ejdmzZ+Pxxx+HUgpHjhzB/Pnz0b1790rPk9xNep7kbrrzJK8g0p0nuZvuPMndLPvb28AH/jiBlpycDKUU/Pz8UKdOHTF5krtJz5PcrSLyiO43u1ulc/36daxcuRIHDhyAk5MTunXrhnHjxsHV1bXS8yR3k54nuZvuPMkriHTnSe6mO09yNzO7e4X//PPPo3HjxujXrx+UUti8eTOuXLmCt956q9LzJHeTnie5m+48ySuIdOdJ7qY7T3I3C+2ngStY3759y7StMvIkd5OeJ7mb7jzJK4h050nupjtPcjczu1ul07BhQxw6dMhy+9ixY2jYsKGIPMndpOdJ7qY7T/IKIt15krvpzpPczZKplH0d0gkPD8eJEyfw6KOPwtHREadOnULNmjXh6upq0xlsnXmSu0nPk9xNd57kFUS68yR3050nuZuZ3Q38Cxcu3PP+Bg0aVFqe5G7S8yR3q4g86SuIuPqq8rMqIs/uBn5+fj727t2L69evF9lu6xv76syT3E16nuRuuvOuX7+O5cuXa11BJDVPcjfdeZK7mdndwP/LX/4CpVSxV1S2vrGvzjzJ3aTnSe6mO0/yCiLdeZK76c6T3M2iXKd8K0F4eLjYPMndpOdJ7qY7T/IKIt15krvpzpPczczuVukEBARg//79MJlM4vIkd5OeJ7mb7jzJK4h050nupjtPcjczu/uXtg899BBGjRpluVyoUgoODg5ITU2t9DzJ3aTnSe6mO+/s2bMYPnx4sRU/PXr0sGkFkeQ8yd2M9lwBOzyG36NHD6xdu9bm9yatyDzJ3aTnSe6mO0/6CiKuvrItT3I3M7t7he/l5QUPDw+ReZK7Sc+T3E13nqenp9YVRJLzJHfTnSe5m5ndDXxvb2+EhYXhiSeegLOzs2W7rasvdOZJ7iY9T3I33XljxowpccWPrT/IkvMkd9OdJ7mbmd0N/O7du5frX5pVZJ7kbtLzJHfTnXf16lVs3bpVS5b0PMnddOdJ7mbmOGfOnDlaEytYixYt4OjoiNOnTyM3Nxdt2rRB165dReRJ7iY9T3I33XmnTp2Co6MjGjRoUOS9Sm0lOU9yN915kruZ2d2yzISEBIwfPx7nz5/HxYsXMWHCBGzatElEnuRu0vMkd9OdZ17x06pVK7Ro0QI+Pj5o0aKFzd0k50nupjtPcjeLcq3irwQRERHqypUrltuZmZnl+scIOvMkd5OeJ7mb7rzAwEB14cIFm7vYU57kbrrzJHczs7tX+CaTCbVq1bLcrl27drn+uqMzT3I36XmSu+nOk7yCSHee5G668yR3M7O7k7bNmzfHggULEBUVBQCIi4uDj4+PiDzJ3aTnSe6mO0/yCiLdeZK76c6T3M3M7v7hVV5enuUKckopBAQEYPz48XB3d6/0PMndpOdJ7qY7b8uWLSVu79+/v03dJOdJ7qY7T3I3M7tbpePg4IBbt25h2rRp6N27N65du4bHH3/c5r9e68yT3E16nuRuuvMkryDSnSe5m+48yd3M7O4Y/syZM7Fjxw7L7e+++w6zZ88WkSe5m/Q8yd1050leQaQ7T3I33XmSu1loPQV8H5T0Jr7leWNfnXmSu0nPk9xNd57kFUS68yR3050nuZuZ3b3CN5lMSE9Pt9zOzMxElSq2Pw2deZK7Sc+T3E13nuQVRLrzJHfTnSe5m5ndrdIZO3Ys+vfvj3bt2kEpheTkZMyYMUNEnuRu0vMkd9OdJ3kFke48yd1050nuZmZ3q3QAIC0tDT/99BOcnJzQunVreHl5icmT3E16nuRuOvMkryDSnSe5m+48yd0synVAiIisdvv2bbVz506l1B/HZePi4pTJZHog8yR3050nuZuZ3R3DJ7J3klcQ6c6T3E13nuRuFuX6dUFEVpO8gkh3nuRuuvMkdzOzm5O277zzzj3vnzBhQqXlSe4mPU9yt4rIA/674sd8DkDXCiKJeZK76c6T3M3MbgY+0YNC8goi3XmSu+nOk9zNzC5X6dxJKYXz58/jkUceEZcnuZv0PMnddORJXUFUEXmSu+nOk9wNsMOBv3btWixZsgS5ubmWbQ8//DD+/e9/V3qe5G7S8yR3q4g8ospgdxdPe/HFFxEbG4vMzEy8++678PHxQX5+PoKDgys9T3I36XmSu1VEHlFlsLtlmXXq1MEjjzyC5s2b45dffsGAAQNw6tQpEXmSu0nPk9ytIvKIKoPdnbStVq0aDhw4gObNm2Pnzp1o3bo1rl+/LiJPcjfpeZK76cqTvoKIq69sy5Pc7W529wp/1qxZ2LVrF7p27Ypr164hJCQEw4cPF5EnuZv0PMndKiKPqDLY3UlbALh9+zaOHz8OR0dHNGvWrNxrU3XmSe4mPU9yt4rIM5O2gqgi8yR3050nsZvdHdLZt28foqOj4eXlBZPJhOvXr2Pp0qXw8/Or9DzJ3aTnSe6mO0/6CiKuvnownysA+7u0Qt++fVVqaqrldnJysurfv7+IPMndpOdJ7qY7LzAwUJ09e1ZNmTJFnTt3Tm3evFlNmTLF5m6S8yR3050nuZuZ3R3Dd3FxKXJN6NatW4vJk9xNep7kbrrzpK8g4uqrB/O5AnZ4SMfPzw8zZszA4MGD4ejoiM8//xwNGjTAwYMHAQDt27evtDzJ3aTnSe6mO0/iCqKKypPcTXee5G5mdnfSdsSIEX96n4ODAz7++ONKy5PcTXqe5G66806cOIG4uDhMmzYNf/vb37B//35MnDgRf/3rX63qZA95krvpzpPczczuBj7Rg0D6CiKuvqr8rIrIs5tDOrNmzcL8+fMxYsSIEt/I19pXbDrzJHeTnie5W0XkAbJXEOnOk9xNd57kbmZ28wr/559/hq+vL77//vsS73/yyScrLU9yN+l5krtVRB4AhIWFYfHixZaTwCkpKZg9ezbi4+OtzpKeJ7mb7jzJ3czsZpWOr68vAKBhw4bYs2cPnnzySdSvXx+bNm1C48aNKzVPcjfpeZK7VUQeIHsFke48yd1050nuZmZ3V8scN24c/P394evri2rVquH69etYuXIlIiMjKz1PcjfpeZK76c5LTU3Fv//9b9StWxeXL1/Ghx9+CADw9PTExYsX0aBBgwcmT3I33XmSu5nZzSEds4iICGzdurXItv79+2PLli2Vnie5m/Q8yd1050leQaQ7T3I33XmSu5nZzUlbM1dXV+zZswdPPfUUACApKQnVqlUTkSe5m/Q8yd10561Zs8bmHvaWJ7mb7jzJ3czs7hV+amoqpk6dioyMDABA/fr1sWjRIjRr1qzS8yR3k54nuZuuPOkriLj6yrY8yd3uZncD3+zq1atwdnaGu7u7uDzJ3aTnSe5W3jzpK4i4+sq2PMnd7mZ3A//o0aNYtWoVsrKycGd1W3/r6cyT3E16nuRuuvPS0tLw8ccfY+rUqTh37hyWL1+OV155BXXr1rWpm+Q8yd1050nuZmZ3x/Cjo6MxZMgQNG3atMS/7lRmnuRu0vMkd9Od9/LLL6Nv374AAG9vb/j7++OVV17B6tWrH7g8yd1050nuZlGua21WgqioKLF5krtJz5PcTXdeeHh4sW39+vV7IPMkd9OdJ7mbmd38wyuzLl26YM2aNTh16hQuXrxo+SMhT3I36XmSu+nOM6/4MdO1gkhinuRuuvMkdzOzu2P4PXr0KLbNwcEBX331VaXnSe4mPU9yN915ElcQVVSe5G668yR3M7O7gU/0oJC0gqii8yR3050nuZvdnLRdvnw5Jk6ciOnTp5d4/xtvvFFpeZK7Sc+T3K0i8gDZK4h050nupjtPcjczuxn4rVq1AlC+NagVlSe5m/Q8yd0qIg+QvYJId57kbrrzJHezKNcp30rwzDPPiM2T3E16nuRuuvMkryDSnSe5m+48yd3M7G6Vzq1bt3Dp0iWReZK7Sc+T3E13nuQVRLrzJHfTnSe5m5ndnbQNDQ3F6dOnUadOHVStWhVKqXKtvtCZJ7mb9DzJ3XTnSV5BpDtPcjfdeZK7Wfa3t4F/4cKFErfbcm1o3XmSu0nPk9ytIvKIKoPdDfyCggKsW7cOBw4cgJOTE5566ilERUXZfFJDZ57kbtLzJHfTlSd9BRFXX9mWJ7nb3exmlY7ZzJkzkZeXh8GDB8NkMuHTTz/FL7/8ghkzZlR6nuRu0vMkd9OVJ30FEVdfVX5WReQVof00cAULDg4ucruwsFCFhoaKyJPcTXqe5G668ySvINKdJ7mb7jzJ3czsbpVO/fr1cebMGcvty5cvw9vbW0Se5G7S8yR3050neQWR7jzJ3XTnSe5mZnfH8EeMGIGUlBT4+/vDyckJhw8fhqenp+Ua0ba8b6SuPMndpOdJ7qY7T/IKIt15krvpzpPczczuBv6fvQuMmbXHvXTmSe4mPU9yN9150lcQcfXVg/lcATsc+ET2TuIKoorKk9xNd57kbmYc+ET3WXR0NPLy8hAZGWlZ8VOvXj2bVxBJzpPcTXee5G4W2k8DE9E9SV5BpDtPcjfdeZK7mdndKh0ieyd5BZHuPMnddOdJ7mbGQzpE95nkFUS68yR3050nuZsZBz7RfSZ5BZHuPMnddOdJ7mbGgU9EZBA8hk9EZBAc+EREBsGBT0RkEBz4REQGwYFPRGQQ/w8ABBJtfNfyBgAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "barplot = sns.barplot(x=pca_labels, y=pca.explained_variance_ratio_)\n", "barplot.set_xticklabels(pca_labels, rotation=90)\n", "plt.title('explained variance ratio')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that already the first three components together account for 50% of the total variance in the dataset. We can visualise this more systematically by plotting the cumulative sum of the explained variance rations:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/Users/leandro/git/dslectures/env/lib/python3.7/site-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variables as keyword args: x, y. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n", " FutureWarning\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sns.lineplot(np.arange(len(pca.explained_variance_ratio_)), np.cumsum(pca.explained_variance_ratio_))\n", "plt.ylabel('total explained variance')\n", "plt.xlabel('number of included PCA components')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We see that with 7 of the 17 axes we already explain 90% percent of the variance in the dataset and that after 10 we almost reach 100%. That means the components 10-17 probably don't contain much information and can be discarded." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### PCA vector composition\n", "Naturally, we would like to know what these principal components mean. For example the first component contains 25% of the total variance. What information is stored in that component? The `pca` object also contains the `components_` which define the transformation between the original dataset and the transformed one. We can print a few of the components and interpret them:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "def print_tabular(labels, values):\n", " df = pd.DataFrame(data=values, columns=labels)\n", " display(df.T)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "The first three principal components:\n" ] }, { "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", "
012
longitude0.096240-0.4404920.279899
latitude-0.0935380.511034-0.231581
housing_median_age-0.228990-0.102574-0.203651
total_rooms0.4808230.1103030.003569
total_bedrooms0.4810680.046098-0.114091
population0.466057-0.002745-0.121771
households0.4829010.023411-0.149648
median_income0.0838290.0661700.381905
rooms_per_household0.0295420.2993680.519937
bedrooms_per_household0.0063540.2045080.343374
bedrooms_per_room-0.029537-0.230823-0.379778
population_per_household-0.002497-0.0036640.003476
ocean_proximity_INLAND-0.0107140.3261730.037222
ocean_proximity_<1H OCEAN0.054867-0.4041360.152350
ocean_proximity_NEAR BAY-0.0673250.233822-0.269103
ocean_proximity_NEAR OCEAN-0.003977-0.076523-0.030817
ocean_proximity_ISLAND-0.006251-0.0090710.001835
\n", "
" ], "text/plain": [ " 0 1 2\n", "longitude 0.096240 -0.440492 0.279899\n", "latitude -0.093538 0.511034 -0.231581\n", "housing_median_age -0.228990 -0.102574 -0.203651\n", "total_rooms 0.480823 0.110303 0.003569\n", "total_bedrooms 0.481068 0.046098 -0.114091\n", "population 0.466057 -0.002745 -0.121771\n", "households 0.482901 0.023411 -0.149648\n", "median_income 0.083829 0.066170 0.381905\n", "rooms_per_household 0.029542 0.299368 0.519937\n", "bedrooms_per_household 0.006354 0.204508 0.343374\n", "bedrooms_per_room -0.029537 -0.230823 -0.379778\n", "population_per_household -0.002497 -0.003664 0.003476\n", "ocean_proximity_INLAND -0.010714 0.326173 0.037222\n", "ocean_proximity_<1H OCEAN 0.054867 -0.404136 0.152350\n", "ocean_proximity_NEAR BAY -0.067325 0.233822 -0.269103\n", "ocean_proximity_NEAR OCEAN -0.003977 -0.076523 -0.030817\n", "ocean_proximity_ISLAND -0.006251 -0.009071 0.001835" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "print('The first three principal components:')\n", "print_tabular(feature_labels, pca.components_[:3])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The way to read this table is to think of it as a recipe to construct the first three principal components coordinates from the original features. Given a new datapoint we can for example calculate the first principal component by multiplying the longitude by the value in column `0` plus the latitude multiplied by the corresponding value in column `0` etc.. This weighted sum of the original featuers will yield the first coordinate in the PCA coordinate system.\n", "\n", "${x_{PC,i}} = \\sum_{i=0}^{n-1} w_{i,j} \\cdot x_{feat, j}$\n", "\n", "From this we can see that the first component mainly consists of the three features `total_rooms`, `total_bedrooms`, and `households`:\n", "\n", "$x_{PC,0} = 0.096 \\cdot longitude + (-0.093) \\cdot lattidue + (-0.229) \\cdot housing\\_median\\_value + \\dots$" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualisation\n", "We can now use the principle components to visualise the data. We have a closer look at the first two components." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sns.scatterplot(x=X_pca[:,0], y=X_pca[:,1], alpha=0.1)\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "jplot = sns.jointplot(x=X_pca[:,0], y=X_pca[:,1], kind=\"hex\", color=\"#4CB391\",xlim=[-5,10], ylim=[-5,10])\n", "jplot.set_axis_labels('Principal component 1', 'Principal component 2')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We observe that there seem to be two visible clusters along these two axis. We could now investigate these clusters further and might find some property of the dataset that might be useful to us or our customer. Let's do it!" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Exercise 1\n", "Create plots of other principal components and investigate if you find any interesting features." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Advanced: Cluster investigation\n", "In this section we have a closer look at the clusters we observed in the previous plot. The clustering seems to mainly happen along the second principle component. Let's plot a fine-grained histogram of that component:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.hist(X_pca[:,1], bins=np.linspace(-5, 5, 50))\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the clusters seem to be seperated around `-0.5`. We can create a mask for the data points that are below that threshold." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "left_group = X_pca[:,1]<-0.5" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Furthermore, we only want to investigate the features that contribute to the second principal component. Therefore, we set another threshold for the minimum absolute feature coefficient." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "pc1_features = list(feature_labels[np.abs(pca.components_[1])>0.2])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "These are the components that mainly contribute to the second principal component:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "['longitude',\n", " 'latitude',\n", " 'rooms_per_household',\n", " 'bedrooms_per_household',\n", " 'bedrooms_per_room',\n", " 'ocean_proximity_INLAND',\n", " 'ocean_proximity_<1H OCEAN',\n", " 'ocean_proximity_NEAR BAY']" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pc1_features" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Unfortunately, the `StandardScaler` transformed the DataFrame `X` into an array `X_std`. Let's transform it to a DataFrame again:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "X_std = pd.DataFrame(data=X_std, columns=feature_labels)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "With `loc` and the mask we just created we can now slice out the datapoints of the left and right group. We have a look at the statistics of the relevant features of the two groups with `.describe()`." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "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", "
longitudelatituderooms_per_householdbedrooms_per_householdbedrooms_per_roomocean_proximity_INLANDocean_proximity_<1H OCEANocean_proximity_NEAR BAY
count8602.0000008602.0000008602.0000008602.0000008602.0000008602.0000008602.0000008602.000000
mean0.767273-0.858163-0.236650-0.0990110.289184-0.6303240.729185-0.345365
std0.3740530.2730840.5230870.2057401.1847090.3891280.8029840.034913
min-1.577731-1.443513-2.051770-1.764901-1.787272-0.704163-0.886320-0.345741
25%0.619362-0.902791-0.611887-0.208570-0.519333-0.7041631.128260-0.345741
50%0.709243-0.804902-0.279346-0.1119080.105280-0.7041631.128260-0.345741
75%0.874025-0.7349810.105771-0.0123030.870705-0.7041631.128260-0.345741
max2.5018711.3673091.9014822.48450513.8565101.4201261.1282602.892336
\n", "
" ], "text/plain": [ " longitude latitude rooms_per_household bedrooms_per_household \\\n", "count 8602.000000 8602.000000 8602.000000 8602.000000 \n", "mean 0.767273 -0.858163 -0.236650 -0.099011 \n", "std 0.374053 0.273084 0.523087 0.205740 \n", "min -1.577731 -1.443513 -2.051770 -1.764901 \n", "25% 0.619362 -0.902791 -0.611887 -0.208570 \n", "50% 0.709243 -0.804902 -0.279346 -0.111908 \n", "75% 0.874025 -0.734981 0.105771 -0.012303 \n", "max 2.501871 1.367309 1.901482 2.484505 \n", "\n", " bedrooms_per_room ocean_proximity_INLAND ocean_proximity_<1H OCEAN \\\n", "count 8602.000000 8602.000000 8602.000000 \n", "mean 0.289184 -0.630324 0.729185 \n", "std 1.184709 0.389128 0.802984 \n", "min -1.787272 -0.704163 -0.886320 \n", "25% -0.519333 -0.704163 1.128260 \n", "50% 0.105280 -0.704163 1.128260 \n", "75% 0.870705 -0.704163 1.128260 \n", "max 13.856510 1.420126 1.128260 \n", "\n", " ocean_proximity_NEAR BAY \n", "count 8602.000000 \n", "mean -0.345365 \n", "std 0.034913 \n", "min -0.345741 \n", "25% -0.345741 \n", "50% -0.345741 \n", "75% -0.345741 \n", "max 2.892336 " ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_std.loc[left_group, pc1_features].describe()" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "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", "
longitudelatituderooms_per_householdbedrooms_per_householdbedrooms_per_roomocean_proximity_INLANDocean_proximity_<1H OCEANocean_proximity_NEAR BAY
count10841.00000010841.00000010841.00000010841.00000010841.00000010841.00000010841.00000010841.000000
mean-0.6088070.6809260.1877750.078563-0.2294590.500143-0.5785860.274036
std0.9191160.8284701.2234491.3214000.7489861.0526360.7247781.273936
min-2.391655-1.406222-1.655426-1.667938-2.026123-0.704163-0.886320-0.345741
25%-1.2981010.299849-0.246291-0.191329-0.724091-0.704163-0.886320-0.345741
50%-0.9435700.9291380.056233-0.092209-0.3093061.420126-0.886320-0.345741
75%-0.0347731.1435620.3869520.0400830.1462781.420126-0.886320-0.345741
max2.6217132.93820058.06979076.7363886.0380551.4201261.1282602.892336
\n", "
" ], "text/plain": [ " longitude latitude rooms_per_household \\\n", "count 10841.000000 10841.000000 10841.000000 \n", "mean -0.608807 0.680926 0.187775 \n", "std 0.919116 0.828470 1.223449 \n", "min -2.391655 -1.406222 -1.655426 \n", "25% -1.298101 0.299849 -0.246291 \n", "50% -0.943570 0.929138 0.056233 \n", "75% -0.034773 1.143562 0.386952 \n", "max 2.621713 2.938200 58.069790 \n", "\n", " bedrooms_per_household bedrooms_per_room ocean_proximity_INLAND \\\n", "count 10841.000000 10841.000000 10841.000000 \n", "mean 0.078563 -0.229459 0.500143 \n", "std 1.321400 0.748986 1.052636 \n", "min -1.667938 -2.026123 -0.704163 \n", "25% -0.191329 -0.724091 -0.704163 \n", "50% -0.092209 -0.309306 1.420126 \n", "75% 0.040083 0.146278 1.420126 \n", "max 76.736388 6.038055 1.420126 \n", "\n", " ocean_proximity_<1H OCEAN ocean_proximity_NEAR BAY \n", "count 10841.000000 10841.000000 \n", "mean -0.578586 0.274036 \n", "std 0.724778 1.273936 \n", "min -0.886320 -0.345741 \n", "25% -0.886320 -0.345741 \n", "50% -0.886320 -0.345741 \n", "75% -0.886320 -0.345741 \n", "max 1.128260 2.892336 " ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_std.loc[~left_group, pc1_features].describe()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "> Note: The `~` operator creates the boolean complement. So `True` becomes `False` and vice versa." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Just looking at the feature mean of the two groups we see that there seems to be a significant difference between almost all features. To have a more detailed view let's compare the histograms of each feature:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl4AAAE/CAYAAACXYc3kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO3de3BUZZ7/8U9fkkAuEMDuQCEDq8yqiIIaGXXXZHQYEghZWHBnkCwZxivsCFOoKbMQBJQUiBmyMC7M1I76Wx284IWAGAMKI44bFKVKLDRaIzflYm5ApDvppG+/P6jJkAFzId1PdyfvV5VV9NOnz/c5XzvwyTlPn7YEg8GgAAAAEHbWSE8AAACgtyB4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwBR7cMPP9TkyZNDvt9FixapsrJSklRUVKT9+/d36fUnT57UFVdcEfJ5AejZCF4AeqXi4mLdcsstkqTKykpxS0MAJtgjPQEA6IwzZ85o2bJl+uKLL2SxWHTrrbfqwQcflN1u1zXXXKP77rtP//d//6eamhrl5+dr9uzZ8vv9WrVqlXbu3KmUlBRde+21OnDggJ5//nnNmjVLeXl5qqqqUk1NjR5++GGtWrVKJSUlysvLU3Z2tiS1bpedna3t27ertLRUffv21ejRo9vM75VXXtGLL76oQCCg1NRULV68WJdffnkkWgUginHGC0BMWL58uVJTU/XGG2/otdde05dffqlnnnlGktTS0qIBAwbopZde0tq1a/Wb3/xGzc3NeuWVV/TZZ59p69ateumll/TNN9+ct98FCxbI6XSqpKREY8aM+d76dXV1WrhwoX7729/q9ddf19ChQ1uf27Nnj8rKyrRhwwaVlZXpnnvu0bx580LfBAAxj+AFICa89957+vd//3dZLBbFx8drxowZeu+991qf/8lPfiJJuvrqq9XS0qLGxkbt2rVLU6ZMUUJCguLj4/Xzn//8ouvv3btX//iP/6iRI0dKUpt9vfvuuzpy5IhmzJihKVOm6Mknn1RDQ4NOnz590fUA9ExcagQQEwKBwHmPfT5f6+OEhARJksVikSQFg0HZ7W3/irNaO/e75rnrvbxeb+t+zx0/d9+BQEBTpkxRQUFB6+Oamhr179+/U/UA9B6c8QIQE/75n/9ZGzZsUDAYVEtLizZu3Ni6OP77ZGZmasuWLWppaZHP59OmTZsuuJ3NZmsNcQMHDmz9hOPXX3+tL7/8UpKUnp6ur776Sl988YUk6fXXX299/T/90z/pzTffVE1NjSTpxRdf1C9+8YvuHTCAHokzXgBiQlFRkZYvX67c3Fx5vV7deuutmjNnTruvmTZtmg4dOqSpU6cqMTFRl156qfr27XveduPHj9eCBQu0fPlyzZ07V4WFhdq1a5cuu+wypaenSzobyEpKSvTwww8rLi5ON954Y+vrb731Vt1777266667ZLFYlJycrKeeeqr17BsA/JUlyGeoAfRQ77//vurr6zVlyhRJZxfoJyQktF4SBADTCF4Aeqzq6moVFhaqvr5efr9fV155pZYuXaqUlJRITw1AL0XwAgAAMITF9QAAAIYQvAAAAAwheAEAABhC8AIAADAkZu7jdeqUW4FAeD8HMGhQsurrXWGt0ZvQz9Cjp6FFP0OPnoYW/Qy9cPfUarVowICk730+ZoJXIBAMe/D6ax2EDv0MPXoaWvQz9OhpaNHP0ItkTzsVvPLz81VfX9/63WSPPfaYvv76a61fv15er1ezZ89WXl6eJKmyslIrVqxQc3OzJk6cqAULFkiSqqqqVFRUJJfLpfT0dC1btuy871EDAADoyTpc4xUMBnXw4EFt3ry59b/BgwertLRUL7zwgjZv3qyXX35ZX331lTwejxYuXKh169apvLxc+/fv165duyRJBQUFWrx4sbZt26ZgMKiNGzeG/eAAAACiSYennA4ePCiLxaJ7771X9fX1+tnPfqakpCTddNNNSk1NlSRlZWWpoqJC48aN0/DhwzVs2DBJUm5urioqKjRy5Eh5PB6NHTtW0tnvT1u7dq1mzpwZxkMDAAAdCQaDcrka1NTkUiDgj/R0wq6mxqpAIBCSfdnt8RowwCGbrfNX8Drc8rvvvtPNN9+spUuXyuPxKD8/XxMnTpTD4Wjdxul06tNPP1VNTc1549XV1eeNOxwOVVdXd3qS0tnFcCY4HHyVSCjRz9Cjp6FFP0OPnoZWuPt55MgRBYMBOZ1DZLPZ+XL3TgoGgzpzpkGNjad02WWXdfp1HQav6667Ttddd50kKTExUXfccYdWrFihOXPmtNnOYrHoQt8+1N54V9TXu8K+GM7hSFFt7Zmw1uhN6Gfo0dPQop+hR09Dy0Q/v/vOpbS0SyVZ5fcHJfXsxfx2u1U+X2jOePXtm6Lq6lNt/h9ZrZZ2TxZ1uMbr448/1u7du1sfB4NBDR06VHV1da1jNTU1cjqdSktL69R4bW2tnE5n548MAACESVAWC7f1vBgXc3aww06fOXNGq1atUnNzs1wulzZt2qQnn3xSu3fv1smTJ9XU1KTt27crIyNDY8aM0aFDh3TkyBH5/X5t3bpVGRkZGjp0qBISErR3715JUllZmTIyMrp+hAAAADGsw0uNt912m/bt26epU6cqEAho5syZuuGGG7RgwQLl5+fL6/Xqjjvu0LXXXitJWrlypebNm6fm5mZlZmYqOztbklRSUqKioiK53W6NGjVK+fn54T0yAABwUfyyqtkX+v0m2CWbOr7M5/P5tGHD/2r79rdksVjk9/s1ceJkzZr1S+Nr0A4c+EpLlizUH/8YmrsxWIIXWoAVhVjjFXvoZ+jR09Cin6FHT0PLRD+//faIBg8e3mas0WdVxZ6673nFxcsed4kS7R0HryeeKNapU/VatGiZUlJS5Ha7tHBhgTIybtP06T/r1hy6ssbrrbe26ve//2/Z7Xa9+uobF9zm7/vX0Rov7mAKAFEqXGcdLqSzZyKAcKupqdb27eXatOktpaSc/URnUlKyHnzwER06dECSVFy8VA0NDTp27BvNnTtfqakDtGZNiVpaWpSamqqCgoW69NJheuCB+3TXXffp+uvTdeLEcc2bd7/Kyt5UcfFSWSwWHTx4QC6XS7Nn363s7Jw283C5XHr//V1aurRYy5cvCdnxEbwAIEo1+xSWsw4XcvZMhJFSQLuqqj7TiBGXqV+/fm3Ghw8foeHDR7Q+7t+/v1atKpXX69Wdd07T44+v1FVXXa2dO9/R0qWL9Ic/PNdundraGv3ud8/o5Ml63X33LN144480aNAlrc8nJyeruPhJnThxPKTHx8cYAABAVDl3Hdef/vSOZs+eqfz8n+uee/62PnzUqNGSpG++OaKUlBRdddXVkqTbbx+vo0e/kcvV/hdhT5qUK7vdLqczTddcM0affvpJGI7kfPx+AwAAosYVV1ylw4cPyu12KSkpWbfdNl633Ta+9VLhXyUkJEj6vi+8DioQ8LcJcD5f2+v2595tPhgMdOnu893BGS8AABA1Bg8eoqysSVq+fKnOnDn7wQK/36/Kyj/Laj0/tvzgB8PV0NCgqqrPJEk7dryttLQh6tevv/r3T21dF/bnP7/b5nU7d76tYDCob789oc8/368xY8aG87BaccYLAAC0kWA/u+4vHPvtjIceKtRLL23Q/Pn3KxiUvN4WXX31NSopWXvetvHx8XrssRVavXqVPJ4m9evXX489tkKSlJeXr+LipXrzzS269dYft3ldc7NHd989S15viwoKFql//9TuHl6ncDuJc/Ax6NCin6FHT0Mr2vsZro/0X0hnP+bfkWjvaayJ1O0kejK73aplyx7VddfdoEmTcru9v67eToJLjQAAAIZwqREAAPQqixYtjVhtzngBAAAYQvACAAAwhOAFAABgCMELAADAEBbXAwCANvpammX1eUK+34C9j5qCCSHfbywheAEAgDasPo+aPtke8v32HTtBsnUcvHw+nzZs+F9t3/6WLBaL/H6/Jk6crFmzftnma4DCqbGxUStWPKavvz6sYDCo/Py7NH58Vrf3S/ACAABR5Te/eUKnTtXrd797VikpKXK7XVq4sEBJScmaPv1nRubwxz/+P6WlDdbjj6/UqVMnNXv2TF1/fboGDhzUrf0SvAAAQNSoqanW9u3l2rTpLaWkpEiSkpKS9eCDj7R+72Jx8VI1NDTo2LFvNHfufKWmDtCaNSVqaWlRamqqCgoW6tJLh+mBB+7TXXfdp+uvT2/9ku2ysjdVXLxUFotFBw8ekMvl0uzZdys7O6fNPK677noNGzZCkjRgwED169dP9fX1BC8AANBzVFV9phEjLlO/fv3ajA8fPkLDh49ofdy/f3+tWlUqr9erO++cpscfX6mrrrpaO3e+o6VLF+kPf3iu3Tq1tTX63e+e0cmT9br77lm68cYfadCgv30/5Y033tT65x07tsvr9eof/uGybh8fn2oEAABR5dx1XH/60zuaPXum8vN/rnvuyW8dHzVqtCTpm2+OKCUlRVdddbUk6fbbx+vo0W/kcrnarTFpUq7sdruczjRdc80YffrpJxfcbufOd7R27W/0+ONPyG7v/vkqzngBAICoccUVV+nw4YNyu11KSkrWbbeN1223jW+9VPhXCQlnF+kHAsEL7CWoQMDfJsD5fL42W9hsf4tAwWCgzeO/evXVl/TCC89r9er/1uWXj+zmkZ3FGS8AABA1Bg8eoqysSVq+fKnOnDkjSfL7/aqs/LOs1vNjyw9+MFwNDQ2qqvpMkrRjx9tKSxuifv36q3//1NZ1YX/+87ttXrdz59sKBoP69tsT+vzz/RozZmyb59977129/PKLWr/+6ZCFLokzXgAA4O8E7H3O3vohDPvVhU5Q/Z2HHirUSy9t0Pz59ysYlLzeFl199TUqKVl73rbx8fF67LEVWr16lTyeJvXr11+PPbZCkpSXl6/i4qV6880tuvXWH7d5XXOzR3ffPUteb4sKChapf//UNs8//fTv1dLi0SOPPNg6VlhYpCuvHNX1Az+HJRgMdqIFkVdf7/qe04mh43CkqLb2TFhr9Cb0M/ToaWhFez8bfVZV7KkzUit73CVKtAe6vZ9o72msMdHPb789osGDh4e1RjSx261atuxRXXfdDZo0Kbfb+/v7/lmtFg0alPy923OpEQAAwBAuNQIAgF5l0aKlEavNGS8AAHo1i4LB7l9m7o0uZrUWwQsAgF4sPr6PTp+uk8/nvagg0VsFg0G53d/Jbo/v0uu41AgAQC82YIBDLleDTp6sViDgj/R0ws5qtSoQCM0ZPrs9XgMGOLr2mpBUBgAAMclisSglJVUpKakdb9wDRPqTt1xqBAAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAM6XTweuKJJ1RYWChJqqqq0vTp05WVlaVFixbJ5/NJko4fP668vDxlZ2dr7ty5crvdkqTvvvtO9913nyZOnKi8vDzV1taG4VAAAACiW6eC1+7du7Vp06bWxwUFBVq8eLG2bdumYDCojRs3SpKWLVummTNnqqKiQqNHj9a6deskSf/1X/+l9PR0vfXWW/q3f/s3FRcXh+FQAAAAoluHwev06dMqLS3VnDlzJEnHjh2Tx+PR2LFjJUnTpk1TRUWFvF6vPvroI2VlZbUZl6R3331Xubm5kqTJkyfrvffek9frDcsBAQAARCt7Rxs8+uijWrBggU6cOCFJqqmpkcPhaH3e4XCourpap06dUnJysux2e5vxv3+N3W5XcnKyTp48qbS0tE5PdNCg5M4fVTc4HClG6vQW9DP06GloRXM/q0+1KCkpwUitPn3i5BgQH5J9RXNPYxH9DL1I9rTd4PXKK69oyJAhuvnmm/X6669LkoLB4HnbWSyW7x3/PlZr19b119e7FAicXyOUHI4U1daeCWuN3oR+hh49Da1o76fHZ5Xb3Wymlser2tru14r2nsYa+hl64e6p1Wpp92RRu8GrvLxctbW1mjJlihoaGtTY2CiLxaK6urrWbWpra+V0OjVw4EC5XC75/X7ZbLbWcUlyOp2qq6vT4MGD5fP55HK5lJqaGqJDBAAAiA3tnnZ69tlntXXrVm3evFnz58/X7bffrhUrVighIUF79+6VJJWVlSkjI0NxcXFKT09XeXl5m3FJyszMVFlZmaSzYS49PV1xcXHhPC4AAICo0+EarwspKSlRUVGR3G63Ro0apfz8fEnSkiVLVFhYqPXr12vIkCFavXq1JOnXv/61CgsLlZOTo5SUFJWUlITuCAAAAGKEJXihxVlRiDVesYd+hh49Da1o72ejz6qKPXUdbxgC2eMuUaI90O39RHtPYw39DL1Ir/HizvUAAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYYo/0BIBo1tfSLKvPY6RWwN5HTcEEI7UAAJFB8ALaYfV51PTJdiO1+o6dINkIXgDQk3GpEQAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwpFPBa82aNZo0aZJycnL07LPPSpIqKyuVm5urCRMmqLS0tHXbqqoqTZ8+XVlZWVq0aJF8Pp8k6fjx48rLy1N2drbmzp0rt9sdhsMBAACIXh0Grz179uiDDz7Qli1b9Nprr+n555/XF198oYULF2rdunUqLy/X/v37tWvXLklSQUGBFi9erG3btikYDGrjxo2SpGXLlmnmzJmqqKjQ6NGjtW7duvAeGQAAQJTpMHiNGzdOzz33nOx2u+rr6+X3+/Xdd99p+PDhGjZsmOx2u3Jzc1VRUaFjx47J4/Fo7NixkqRp06apoqJCXq9XH330kbKystqMAwAA9Cb2zmwUFxentWvX6plnnlF2drZqamrkcDhan3c6naqurj5v3OFwqLq6WqdOnVJycrLsdnub8a4YNCi5S9tfLIcjxUid3iLW++k/3ShLYoKRWvF94pSY2nG/Yr2n0Saa+1l9qkVJSWbef336xMkxID4k+4rmnsYi+hl6kexpp4KXJM2fP1/33nuv5syZo8OHD5/3vMViUTAY7NJ4V9TXuxQInL+fUHI4UlRbeyasNXqTntDPJL9XTY3NRmoFPV65O+hXT+hpNIn2fnp8VrndZt5/Ho9XtbXdrxXtPY019DP0wt1Tq9XS7smiDi81HjhwQFVVVZKkvn37asKECfrwww9VV1fXuk1NTY2cTqfS0tLajNfW1srpdGrgwIFyuVzy+/1txgEAAHqTDoPX0aNHVVRUpJaWFrW0tGjHjh2aMWOGDh06pCNHjsjv92vr1q3KyMjQ0KFDlZCQoL1790qSysrKlJGRobi4OKWnp6u8vLzNOAAAQG/S4aXGzMxM7du3T1OnTpXNZtOECROUk5OjgQMHat68eWpublZmZqays7MlSSUlJSoqKpLb7daoUaOUn58vSVqyZIkKCwu1fv16DRkyRKtXrw7vkQEAAEQZS/BCC7CiEGu8Yk9P6GeSv0FNn2w3Uqvv2Aly2/q3u01P6Gk0ifZ+NvqsqthT1/GGIZA97hIl2gPd3k+09zTW0M/Qi/o1XgAAAAgNghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIZ0Kng99dRTysnJUU5OjlatWiVJqqysVG5uriZMmKDS0tLWbauqqjR9+nRlZWVp0aJF8vl8kqTjx48rLy9P2dnZmjt3rtxudxgOBwAAIHp1GLwqKyv1/vvva9OmTSorK9Nnn32mrVu3auHChVq3bp3Ky8u1f/9+7dq1S5JUUFCgxYsXa9u2bQoGg9q4caMkadmyZZo5c6YqKio0evRorVu3LrxHBgAAEGU6DF4Oh0OFhYWKj49XXFycLr/8ch0+fFjDhw/XsGHDZLfblZubq4qKCh07dkwej0djx46VJE2bNk0VFRXyer366KOPlJWV1WYcAACgN7F3tMEPf/jD1j8fPnxY5eXlmjVrlhwOR+u40+lUdXW1ampq2ow7HA5VV1fr1KlTSk5Olt1ubzPeFYMGJXdp+4vlcKQYqdNbxHo//acbZUlMMFIrvk+cElM77les9zTaRHM/q0+1KCnJzPuvT584OQbEh2Rf0dzTWEQ/Qy+SPe0weP3VX/7yF91///165JFHZLfbdejQoTbPWywWBYPB817X3nhX1Ne7FAicv59QcjhSVFt7Jqw1epOe0M8kv1dNjc1GagU9Xrk76FdP6Gk0ifZ+enxWud1m3n8ej1e1td2vFe09jTX0M/TC3VOr1dLuyaJOLa7fu3evZs+erYceekj/+q//qrS0NNXV1bU+X1NTI6fTed54bW2tnE6nBg4cKJfLJb/f32YcAACgN+kweJ04cUK/+tWvVFJSopycHEnSmDFjdOjQIR05ckR+v19bt25VRkaGhg4dqoSEBO3du1eSVFZWpoyMDMXFxSk9PV3l5eVtxgEAAHqTDi81Pv3002pubtbKlStbx2bMmKGVK1dq3rx5am5uVmZmprKzsyVJJSUlKioqktvt1qhRo5Sfny9JWrJkiQoLC7V+/XoNGTJEq1evDtMhAQAARCdL8EILsKIQa7xiT0/oZ5K/QU2fbDdSq+/YCXLb+re7TU/oaTSJ9n42+qyq2FPX8YYhkD3uEiXaA93eT7T3NNbQz9CLiTVeAAAA6D6CFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACG2CM9AQBn2WwWJfkb2t3Gf7pRSX5vt2sF7H3UFEzo9n4AAF1D8AKihMXXrKZ977a/TWKCmhqbu12r79gJko3gBQCmcakRAADAEIIXAACAIQQvAAAAQwheAAAAhrC4HgBglF9WNfvCXyfBLtkUCH8hoAsIXgAAo5p9UsWeurDXyR53iRL5Vw5RhkuNAAAAhvC7AABAVptVjSG4/Fd9qkUeX/u/0we7XwaIWQQvAIBavAG983H3L/8lJSXI7W7/Jr/j0y/pdh0gVnGpEQAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEHukJwB0VV9Ls6w+j5FaNovfSB0AQO9A8ELMsfo8avpku5FaKWN+bKQOAKB34FIjAACAIQQvAAAAQzodvFwulyZPnqyjR49KkiorK5Wbm6sJEyaotLS0dbuqqipNnz5dWVlZWrRokXw+nyTp+PHjysvLU3Z2tubOnSu32x3iQwEAAIhunQpe+/bt05133qnDhw9LkjwejxYuXKh169apvLxc+/fv165duyRJBQUFWrx4sbZt26ZgMKiNGzdKkpYtW6aZM2eqoqJCo0eP1rp168JzRAAAAFGqU8Fr48aNWrJkiZxOpyTp008/1fDhwzVs2DDZ7Xbl5uaqoqJCx44dk8fj0dixYyVJ06ZNU0VFhbxerz766CNlZWW1GQcAAOhNOvWpxuLi4jaPa2pq5HA4Wh87nU5VV1efN+5wOFRdXa1Tp04pOTlZdru9zXhXDBqU3KXtL5bDkWKkTm8Rjn76TzfKkpgQ8v1eiNVmVWKU1QrFfOL7xCkxlfe6FN0/89WnWpSUZOb9Z7NZQ1aro/2EslZ7+vSJk2NAfNjrhFs0v0djVSR7elG3kwgGg+eNWSyWLo93RX29S4HA+fsJJYcjRbW1Z8JaozcJVz+T/F41NTaHfL8XkuIPqDGKaiUmJoRkPkGPV27e61H/M+/xWeV2m3n/+f2BkNRKSkrocD+hqtURj8er2loz/QuXaH+PxqJw99RqtbR7suiiPtWYlpamurq61sc1NTVyOp3njdfW1srpdGrgwIFyuVzy+/1txgEAAHqTiwpeY8aM0aFDh3TkyBH5/X5t3bpVGRkZGjp0qBISErR3715JUllZmTIyMhQXF6f09HSVl5e3GQcAAOhNLupSY0JCglauXKl58+apublZmZmZys7OliSVlJSoqKhIbrdbo0aNUn5+viRpyZIlKiws1Pr16zVkyBCtXr06dEcBAAAQA7oUvHbu3Nn655tvvllbtmw5b5srr7xSr7766nnjQ4cO1fPPP38RUwQAAOgZuHM9AACAIQQvAAAAQwheAAAAhhC8AAAADLmoTzUCQLTpa2mW1efp0mv8pxuV5Pd2uVbA3kdNQTN3lAfQsxC8APQIVp9HTZ9s79JrLIkJF/UtCH3HTpBsBC8AXcelRgAAAEM44wX0QjabRUn+hrDX4ZJc7EiN8yjjMl+392O3B+Xz+TusBfRWBC+gF7L4mtW0792w1+GSXOyw+Tyq/6Ci2/uJi7PL620/wA3JmdrtOkCs4lIjAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAj38QIAGBVvV0hu1tqRfjaPfIoPex2gKwheAACjLL6WkNysteyasqUAAAfmSURBVCNDJ0+Vz07wQnThUiMAAIAhBC8AAABDCF4AAACGsMYLALrIZrMoyd8Q9jp2W9+w1wBgFsELALrI4mtW0753w14nbkxW2GsAMIvgBSBsTJ0ZkiSbxW+kDgB0B8ELQNiYOjMkSSljfmykDgB0B4vrAQAADCF4AQAAGMKlRgCIUnE2M1+tI0nxVouROkBvR/ACgChl8Zv5ah1JGpEzyUgdoLfjUiMAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYwqcaAaCLgrLIG+D2CwC6juAFAF0UlPTVscaw17lhbNhLADCMS40AAACGELwAAAAMIXgBAAAYwhqvHq6vpVlWn8dIrYC9j5qCCUZqAQAQiwhePZzV51HTJ9uN1Oo7doJkI3gBiA5xNsnubzBSi1880VkELwA9wsXc4sHt8cvHbSF6LIu/RY37/mSkFr94orMIXgB6hIu5xUNcnF1er6/LtbjNA4CLRfBCyNhsFiWdc1rff7pRSX5v6OtY/CHfJwB0R0AWNfpC/3m16lMt8vzdfhPskk2BkNeCGUaD1xtvvKH169fL6/Vq9uzZysvLM1keYWbxNatp37t/e5yYoKbG5pDXSRnz45DvE+HBHd7RWwSCUsWeupDvNykpQW53279Hs8ddokROm8QsY//rqqurVVpaqtdff13x8fGaMWOGfvSjH2nkyJGmpgB0mcngEJS5gGLquEzd4V3i8h8uzNTPLzkInWXsvVJZWambbrpJqampkqSsrCxVVFTogQce6NTrrVYzPzym6phikU3WvslmitnsbWpZEuJlDcaFvU5Y2ez6psFmpNQAm12BhJR2t2mSVYGE+O4XM3RcqVa7EpLbP6ZQsdi6XsseZ5PV2/VL1xdT62KYqhPKWp3pqbHjspr7+R0dZ9NPruj6esGO2Ow++f9ut6lxLfJZwvB36wUEZFWzgdUdCTbJavDyaTj/re9o35ZgMBgMW/Vz/P73v1djY6MWLFggSXrllVf06aef6vHHHzdRHgAAIOKM3bn+QvnOYulZZ5cAAADaYyx4paWlqa7ubwsPa2pq5HQ6TZUHAACIOGPB65ZbbtHu3bt18uRJNTU1afv27crIyDBVHgAAIOKMLa5PS0vTggULlJ+fL6/XqzvuuEPXXnutqfIAAAARZ2xxPQAAQG9n7FIjAABAb0fwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEEr3N8/PHHmjZtmnJzczVnzhw1NDREekoxb+/evZo+fbqmTJmiX/ziFzp27Fikp9QjrFmzRr/97W8jPY2Y9sYbb2jSpEn66U9/qg0bNkR6Oj2Cy+XS5MmTdfTo0UhPpUd46qmnlJOTo5ycHK1atSrS04l5a9as0aRJk5STk6Nnn302YvMgeJ3jP//zP7Vq1Sq98cYbGjlypJ5++ulITynmFRQUqLi4WJs3b1Zubq6WL18e6SnFtDNnzmjhwoV65plnIj2VmFZdXa3S0lK98MIL2rx5s15++WV99dVXkZ5WTNu3b5/uvPNOHT58ONJT6REqKyv1/vvva9OmTSorK9Nnn32mt99+O9LTill79uzRBx98oC1btui1117T888/r4MHD0ZkLgSvc5SXl2vkyJHyer2qrq5Wv379Ij2lmNbS0qJf//rXuvLKKyVJV1xxhU6cOBHhWcW2HTt2aMSIEfrlL38Z6anEtMrKSt10001KTU1VYmKisrKyVFFREelpxbSNGzdqyZIlfBVciDgcDhUWFio+Pl5xcXG6/PLLdfz48UhPK2aNGzdOzz33nOx2u+rr6+X3+5WYmBiRuRC8zhEXF6cvv/xSmZmZ+vDDD5WTkxPpKcW0+Ph4TZkyRZIUCAT01FNPafz48RGeVWybOnWq7rvvPtlstkhPJabV1NTI4XC0PnY6naquro7gjGJfcXGx0tPTIz2NHuOHP/yhxo4dK0k6fPiwysvLlZmZGeFZxba4uDitXbtWOTk5uvnmm5WWlhaRefTK4PXWW28pIyOjzX+zZ8+WdPasTGVlpf7jP/5DCxYsiOxEY0h7PW1padHDDz8sn8+n+++/P7ITjRHt9RPdd6Ev7LBYLBGYCdC+v/zlL7rrrrv0yCOPaMSIEZGeTsybP3++du/erRMnTmjjxo0RmYOx72qMJhMnTtTEiRPbjDU3N+udd95pPSPzL//yL3riiSciMb2YdKGeSpLb7dbcuXOVmpqq9evXKy4uLgKziz3f10+ERlpamj7++OPWxzU1NVwiQ9TZu3ev5s+fr4ULF3IFppsOHDiglpYWXXXVVerbt68mTJigL7/8MiJz6ZVnvC7Ebrdr2bJl2r9/v6SzZxyuv/76CM8q9hUUFGj48OFas2aN4uPjIz0dQJJ0yy23aPfu3Tp58qSampq0fft2ZWRkRHpaQKsTJ07oV7/6lUpKSghdIXD06FEVFRWppaVFLS0t2rFjh2644YaIzKVXnvG6EJvNptLSUj366KPy+/1KS0tTcXFxpKcV0z7//HPt2LFDI0eO1NSpUyWdXUvzP//zPxGeGXq7tLQ0LViwQPn5+fJ6vbrjjjt07bXXRnpaQKunn35azc3NWrlyZevYjBkzdOedd0ZwVrErMzNT+/bt09SpU2Wz2TRhwoSIBVpL8EKLHQAAABByXGoEAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADPn/Ya7QIeXLY1oAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl4AAAE/CAYAAACXYc3kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO3df3SU5Z3//9f8SAITAkGciRxk40dxC1QlKiL0nE10v5IEQgqCpx8gh5StH0VXwaLNigHkh6LIZqGwCu121a1oPyVLa0A2DSBUPtWgCOcUTizrWiRYfnQmRIpkwkzm1/cPjtEYmsmPmWsy5Pk4x3OYa6653td95RZeue977tsSiUQiAgAAQNxZEz0BAACAvoLgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAknQ8++EBTpkzpsM+LL76ot99+W5K0fv16VVVVtWvviltvvVUnT57s+mQB4GvsiZ4AAMTDBx98oBEjRkiSHnvsscu2A4BpBC8ASev48eNauXKlmpub5fF4NHLkSP34xz/W1q1bVVdXpzVr1shms2nPnj268cYb1a9fv8u233///ZKkRYsWtb4+ePCgnnnmGVksFt18880Kh8Otdffu3atNmzYpEAioX79+evLJJ3XrrbcmahkAJBFONQJIWpWVlZo2bZq2bNmiXbt26eTJk3rnnXdUUlKim266Sf/0T/+kiRMntvb/a+3f1NLSoscee0yLFi1SVVWV7rzzTvl8PklSfX291q1bp3/7t39TVVWVnnnmGc2fP1/Nzc1x314AyY8jXgCSVllZmd577z397Gc/U319vTweT0wC0P/8z//IbrdrwoQJkqQpU6bo6aefliS999578ng8mjt3bmt/i8Wizz77TCNHjuxxbQBXNoIXgKT1+OOPKxQKadKkSbrrrrt05swZdeXxsxaLpU3/QCBw2XZJstsv/XUZDoc1YcIE/fjHP25978yZM3K5XD3ZFAB9BKcaASStd999V4888ogmT54si8Wiw4cPKxQKSZJsNpuCwWC7z3y9ffDgwaqrq5Mkff755zp48KAk6W//9m8ViUS0b98+SdKePXt0/vx5SdL48eP13nvv6dixY5Kkffv26bvf/a78fn98NxbAFYEjXgCS1sKFC/XII49o0KBB6t+/v+644w599tlnkqS7775bL7zwQutRrC99vX3OnDn60Y9+pIKCAl177bUaN26cJCklJUUvvfSSli9frrVr12rUqFEaMmSIJOnGG2/UypUr9fjjjysSichut2vTpk1yOBxmNx5AUrJEunJcHgAAAN3GqUYAAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgSNLcx+vcOa/C4fje+WLIkAFqbGyKa42+hPWMPdY0tljP2GNNY4v1jL14r6nVatHgwel/9f2kCV7hcCTuwevLOogd1jP2WNPYYj1jjzWNLdYz9hK5ppxqBAAAMITgBQAAYEjSnGoEAACxFwoFde5cg4LBlkRPxQiPx6pwOByTsez2VA0e7JTN1vk4RfACAKAPO3euQf36OZSefo0sFkuipxN3drtVwWDPg1ckEpHX+4XOnWvQ1VcP7fTnONUIAEAfFgy2KD19YJ8IXbFksViUnj6wy0cKCV4AAPRxhK7u6c66EbwAAAAM4RovAADQRkhW+YOxHzfNLtnU8fVVwWBQb7zxc+3a9RtZLBaFQiFNmjRFc+b8g/Ejc8eO/VHLlpXr9dcrYzYmwQsAALThD0o1B87GfNzCcVfLESV5/Mu/vKBz5xr1k5+8qoyMDHm9TSovL1N6+gDNmPG9mM/pr/nNb3bopz99SXZ7bKMSwQvoY+L1m+zldOa3WwD4ksfj1q5d1Xrzzd8oIyNDkpSePkCPP/6kjh8/JklatWq5zp8/r1On/qSHH16gzMzBWr++Qi0tLcrMzFRZWbmuvXa4Hn30Qf3gBw/qttvG6syZ05o/f562bn1LK1cukyR9+ukxNTU1ae7c+1VYWNRmHk1NTXr33X1avnyVnn12WUy3keAF9DHx+k32cjrz2y0AfOno0Y903XXXa+DAgW3as7OvU3b2da2vBw0apDVr1ikQCGjWrOl65pnVGjXq29q7920tX75Y//7vr3VYp6HBo5/85BV9/nmj7r9/ju64404NGXJ16/sDBgzQqlX/rDNnTsd0+yQurgcAAL3I16/j+u1v39bcubNVWvq/9X/+T2lr++jRN0mS/vSnE8rIyNCoUd+WJP3939+jkyf/pKamjh+CPXlysex2u1yuLN188xgdOfL7OGzJ5fG7KAAA6BW+9a1Rqq//VF5vk9LTB+juu+/R3Xff03qq8EtpaWmS/trDriMKh0NtAlww2Pb6iq/faT4SCXfpzvM9xREvAADQK1xzzVAVFEzWs88u14ULFyRJoVBItbW/k9XaPrL8zd9k6/z58zp69CNJ0p49u5WVNVQDBw7SoEGZrdeF/e5377T53N69uxWJRPTnP5/RH/5QpzFjcuK5WW1wxAsAALSRZr90jWY8xo3miScW6Ze/fEMLFsxTJCIFAi369rdvVkXFhnZ9U1NTtXLl81q7do18vosaOHCQVq58XpJUUlKqVauW67/+a7v+7u/uavM5v9+n+++fo0CgRWVlizVoUGYsNq9TLJFI5HLH6Xqdxsamv3JIMXaczgw1NFyIa42+hPWMvVisaXPQavji+t77rUb20dhjTWPLxHr++c8ndM012XGt0Zs899wK5eTcpsmTi2My3jfXz2q1aMiQAX+1P6caAQAADOFUIwAA6DOefnqFgsHEHYnniBcAAIAhBC8AAABDOhW81q9fr8mTJ6uoqEivvvqqJOmpp55Sfn6+pk6dqqlTp2r37t2SpNraWhUXFys/P1/r1q1rHePo0aOaMWOGCgoKtHjx4nb31AAAALjSRb3G68CBA3r//fe1fft2BYNBTZ48WXl5eaqrq9Prr78ul8vV2tfn86m8vFybN2/W0KFDNW/ePO3bt095eXkqKyvTs88+q5ycHJWXl6uyslKzZ8+O68YBAAD0JlGD17hx4/Taa6/JbrfL7XYrFAopLS1Np0+f1tKlS3X69GlNnDhRjz76qI4cOaLs7GwNHz5cklRcXKyamhqNGDFCPp9POTmXblA2ffp0bdiwgeAFAEAv1N/ilzXoi/m4YXs/XYykxXzcZNKpbzWmpKRow4YNeuWVV1RYWKhQKKTx48dr5cqVcjgcmjdvnrZu3SqHwyGn09n6OZfLJbfbLY/H06bd6XTK7XZ3aaId3RMjlpzODCN1+grWM/Z6uqbucy1KTzfzF1+/filyDk41Uqu72EdjjzWNrXivp8djld3e9sojW8Av35HdMa/VLydf9pT+HfYJBoPavPk/VFNTLYvFonA4rMmTp+j73/9Bm8cA9cQ3t/ebmpubtWrVCtXXH5ckzZ17vyZOLLhsX6vV2qWfUadvJ7FgwQI98MADeuihh7R//3699NJLre/NmTNHVVVVKiwsbPc5i8Wiy92jtauLxw1Ukw/rGXuxWFNf0Cqv1x+jGUWp5QuoocFMre5gH/1KrI5wpPVLkd8X6LAPRz06z8Q+Gg6H291eIRKOxOXf3Eg4EvVWDi+88LzOnWvUT37yqjIyMuT1Nqm8vEz9+6drxozv9XgOdrs16hz+4z9ekdOZpRUrnte5c59r7tzZGjPmNl111ZB2fcPhcJufUbQbqEYNXseOHVNLS4tGjRql/v37Kz8/X9XV1crMzFRBwaX0F4lEZLfblZWVpbNnv7ojtsfjkcvlatfe0NDQ5towAEBiWYM+Xfz9rh6PY3Gk6WJzx2G7f06+ZCN4oT2Px61du6r15pu/UUbGpaNI6ekD9PjjT7Y+d3HVquU6f/68Tp36kx5+eIEyMwdr/foKtbS0KDMzU2Vl5br22uF69NEH9YMfPKjbbhvb+pDtrVvf0sqVyyRJn356TE1NTZo7934VFha1mcett96m4cOvkyQNHnyVBg4cqMbGxssGr66K+q3GkydPasmSJWppaVFLS4v27NmjO+64Q88995zOnz+vQCCgLVu2aOLEiRozZoyOHz+uEydOKBQKaceOHcrNzdWwYcOUlpamQ4cOSZKqqqqUm5vb48kDAIArx9GjH+m6667XwIED27RnZ1+nu+76/1pfDxo0SG+8sVV33jlBy5eX6/HH/0k///n/1dSpM7R8+eKodRoaPPrJT17Rhg2b9NJL69XY2PYxanfcMV7XXHONJGnPnl0KBAL6X//r+hhsYSeOeOXl5enw4cOaNm2abDab8vPz9eijj2rw4MGaNWuWgsGg8vPzNWXKFEnS6tWrNX/+fPn9fuXl5bWefqyoqNCSJUvk9Xo1evRolZaWxmQDAADAlePrlyL99rdv6+c/f0XhcEipqWn6939/TZI0evRNkqQ//emEMjIyNGrUtyVJf//392jNmlVqamrqsMbkycWy2+1yubJ0881jdOTI73X33fe067d379vasOFfVFHxr7LbY/Own06NsmDBAi1YsKBNW0lJiUpKStr1nTBhgrZv396ufeTIkdq6dWs3pwkAAK503/rWKNXXfyqvt0np6QN099336O6772k9VfiltLRLp6ovfx1aROFwqE2A++a9Q222r+JPJBJu8/pLW7f+Ur/4xWatXfuSbrhhRA+37CvcuR4AAPQK11wzVAUFk/Xss8t14cKlC9ZDoZBqa38nq7V9ZPmbv8nW+fPndfToR5KkPXt2KytrqAYOHKRBgzJbrwv73e/eafO5vXt3KxKJ6M9/PqM//KFOY8bktHn///2/d7Rly//Vpk0vxzR0STwkGwAAfEPY3u/SlyDiMK6ifFnyiScW6Ze/fEMLFsxTJCIFAi369rdvVkXFhnZ9U1NTtXLl81q7do18vosaOHCQVq58XpJUUlKqVauW67/+a7v+7u/uavM5v9+n+++fo0CgRWVlizVoUGab919++adqafHpyScfb21btGiJRo4c3b0N/xpL5HL3euiFuJ1E8mE9Yy8Wa9octKrmwNnoHWOgcNzVctg7/tp2IrGPfiU9dD4m32p0ONLU3IlvNXptg3pcqy8wsY/++c8ndM012XGt0Zs899wK5eTcpsmTi2My3jfXL9rtJDjVCAAAYAinGgEAQJ/x9NMrot5ANZ444gUAQB+XJFcd9TrdWTeCFwAAfZjdniqv9wvCVxdFIhF5vV/Ibu/a82g51QgAQB82eLBT5841qKnpL4meihFWq1XhcGxONdrtqRo82Nm1z8SkMgAASEo2m11XXz000dMwJtHfZuZUIwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQbqAK9BIhWeUPdtzHfa5FvmDPfl/ioSAAkDgEL6CX8AelmgNnO+yTnp4mr9ffozr3jL26R58HAHQfpxoBAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQzoVvNavX6/JkyerqKhIr776qiSptrZWxcXFys/P17p161r7Hj16VDNmzFBBQYEWL16sYPDSHSFPnz6tkpISFRYW6uGHH5bX643D5gAAAPReUYPXgQMH9P7772v79u361a9+pc2bN+u///u/VV5ero0bN6q6ulp1dXXat2+fJKmsrExLly7Vzp07FYlEVFlZKUlasWKFZs+erZqaGt10003auHFjfLcMAACgl4kavMaNG6fXXntNdrtdjY2NCoVC+uKLL5Sdna3hw4fLbreruLhYNTU1OnXqlHw+n3JyciRJ06dPV01NjQKBgD788EMVFBS0aQcAAOhLOvXIoJSUFG3YsEGvvPKKCgsL5fF45HQ6W993uVxyu93t2p1Op9xut86dO6cBAwbIbre3ae+KIUMGdKl/dzmdGUbq9BWsZ+e5z7UoPT0tar/O9OmIzWbt8Rid1a9fipyDU43U6i720UtCf2mWxRGb/cIRZZzUfilyZLLuncU+GnuJXNNOP6txwYIFeuCBB/TQQw+pvr6+3fsWi0WRSPvH73bU3hWNjU0Kh+P7eF+nM0MNDRfiWqMvYT27xhe0Rn0OYyye1RgKhXs8Rmf5fAE1NJip1R3so19JDwV0sbnnPyuHI03NUcaJ+ALysu6dwj4ae/FeU6vV0uHBoqinGo8dO6ajR49Kkvr376/8/Hx98MEHOnv2q4f5ejweuVwuZWVltWlvaGiQy+XSVVddpaamJoVCoTbtAAAAfUnU4HXy5EktWbJELS0tamlp0Z49ezRz5kwdP35cJ06cUCgU0o4dO5Sbm6thw4YpLS1Nhw4dkiRVVVUpNzdXKSkpGjt2rKqrq9u0AwAA9CVRTzXm5eXp8OHDmjZtmmw2m/Lz81VUVKSrrrpK8+fPl9/vV15engoLCyVJFRUVWrJkibxer0aPHq3S0lJJ0rJly7Ro0SJt2rRJQ4cO1dq1a+O7ZQAAAL2MJXK5C7B6Ia7xSj6sZ9c0B62qOXC2wz6xuMbrnrFX6+2DHdeJlcJxV8thDxup1R3so19JD53Xxd/v6vE4nbnGq39Ovry2QT2u1Rewj8Zer7/GCwAAALFB8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCkU8HrxRdfVFFRkYqKirRmzRpJ0lNPPaX8/HxNnTpVU6dO1e7duyVJtbW1Ki4uVn5+vtatW9c6xtGjRzVjxgwVFBRo8eLFCgaDcdgcAACA3itq8KqtrdW7776rN998U1VVVfroo4+0e/du1dXV6fXXX9e2bdu0bds2TZw4UT6fT+Xl5dq4caOqq6tVV1enffv2SZLKysq0dOlS7dy5U5FIRJWVlXHfOAAAgN4kavByOp1atGiRUlNTlZKSohtuuEGnT5/W6dOntXTpUhUXF2vDhg0Kh8M6cuSIsrOzNXz4cNntdhUXF6umpkanTp2Sz+dTTk6OJGn69OmqqamJ+8YBAAD0JvZoHW688cbWP9fX16u6ulq/+MUvdODAAa1cuVIOh0Pz5s3T1q1b5XA45HQ6W/u7XC653W55PJ427U6nU263u0sTHTJkQJf6d5fTmWGkTl/Benae+1yL0tPTovbrTJ+O2GzWHo/RWf36pcg5ONVIre5iH70k9JdmWRyx2S8cUcZJ7ZciRybr3lnso7GXyDWNGry+9Mknn2jevHl68skndf311+ull15qfW/OnDmqqqpSYWFhu89ZLBZFIpHLtndFY2OTwuH248SS05mhhoYLca3Rl7CeXeMLWuX1+jvsk56eFrVPNKFQuMdjdJbPF1BDg5la3cE++pX0UEAXm3v+s3I40tQcZZyILyAv694p7KOxF+81tVotHR4s6tTF9YcOHdLcuXP1xBNP6N5779XHH3+snTt3tr4fiURkt9uVlZWls2fPtrZ7PB65XK527Q0NDXK5XN3ZHgAAgKQVNXidOXNGjzzyiCoqKlRUVCTpUtB67rnndP78eQUCAW3ZskUTJ07UmDFjdPz4cZ04cUKhUEg7duxQbm6uhg0bprS0NB06dEiSVFVVpdzc3PhuGQAAQC8T9VTjyy+/LL/fr9WrV7e2zZw5Uw8++KBmzZqlYDCo/Px8TZkyRZK0evVqzZ8/X36/X3l5ea2nHysqKrRkyRJ5vV6NHj1apaWlcdokAACA3skSudwFWL0Q13glH9aza5qDVtUcONthn1hc43XP2Kv19sGO68RK4bir5bCHjdTqDvbRr6SHzuvi73f1eJzOXOPVPydfXtugHtfqC9hHYy8prvECAABAzxG8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABDCF4AAACGELwAAAAMIXgBAAAYQvACAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADOlU8HrxxRdVVFSkoqIirVmzRpJUW1ur4uJi5efna926da19jx49qhkzZqigoECLFy9WMBiUJJ0+fVolJSUqLCzUww8/LK/XG4fNAQAA6L2iBq/a2lq9++67evPNN1VVVaWPPvpIO3bsUHl5uTZu3Kjq6mrV1dVp3759kqSysjItXbpUO3fuVCQSUWVlpSRpxYoVmj17tmpqanTTTTdp48aN8d0yAACAXiZq8HI6nVq0aJFSU1OVkpKiG264QfX19crOztbw4cNlt9tVXFysmpoanTp1Sj6fTzk5OZKk6dOnq6amRoFAQB9++KEKCgratAMAAPQl9mgdbrzxxtY/19fXq7q6WnPmzJHT6Wxtd7lccrvd8ng8bdqdTqfcbrfOnTunAQMGyG63t2nviiFDBnSpf3c5nRlG6vQVrGfnuc+1KD09LWq/zvTpiM1m7fEYndWvX4qcg1ON1Oou9tFLQn9plsURm/3CEWWc1H4pcmSy7p3FPhp7iVzTqMHrS5988onmzZunJ598Una7XcePH2/zvsViUSQSafe5jtq7orGxSeFw+3FiyenMUEPDhbjW6EtYz67xBa3yev0d9klPT4vaJ5pQKNzjMTrL5wuoocFMre5gH/1Keiigi809/1k5HGlqjjJOxBeQl3XvFPbR2Iv3mlqtlg4PFnXq4vpDhw5p7ty5euKJJ3TvvfcqKytLZ8+ebX3f4/HI5XK1a29oaJDL5dJVV12lpqYmhUKhNu0AAAB9SdTgdebMGT3yyCOqqKhQUVGRJGnMmDE6fvy4Tpw4oVAopB07dig3N1fDhg1TWlqaDh06JEmqqqpSbm6uUlJSNHbsWFVXV7dpBwAA6Euinmp8+eWX5ff7tXr16ta2mTNnavXq1Zo/f778fr/y8vJUWFgoSaqoqNCSJUvk9Xo1evRolZaWSpKWLVumRYsWadOmTRo6dKjWrl0bp00CAADonSyRy12A1QtxjVfyYT27pjloVc2Bsx32icU1XveMvVpvH+y4TqwUjrtaDnvYSK3uYB/9SnrovC7+flePx+nMNV79c/LltQ3qca2+gH009pLiGi8AAAD0HMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAhBC8AAABD7ImeAACgb7HZLEoPnY97nbC9ny5G0uJeB+gKghcAwChL0K+Lh9+Je53+OfmSjeCF3oVTjQAAAIYQvAAAAAwheAEAABhC8AIAADCk08GrqalJU6ZM0cmTJyVJTz31lPLz8zV16lRNnTpVu3fvliTV1taquLhY+fn5WrduXevnjx49qhkzZqigoECLFy9WMBiM8aYAAAD0bp0KXocPH9asWbNUX1/f2lZXV6fXX39d27Zt07Zt2zRx4kT5fD6Vl5dr48aNqq6uVl1dnfbt2ydJKisr09KlS7Vz505FIhFVVlbGZYMAAAB6q04Fr8rKSi1btkwul0uS1NzcrNOnT2vp0qUqLi7Whg0bFA6HdeTIEWVnZ2v48OGy2+0qLi5WTU2NTp06JZ/Pp5ycHEnS9OnTVVNTE7+tAgAA6IU6dR+vVatWtXnd2Nio8ePHa+XKlXI4HJo3b562bt0qh8Mhp9PZ2s/lcsntdsvj8bRpdzqdcrvdXZrokCEDutS/u5zODCN1+grWs/Pc51qUnh79nkOd6dMRm83a4zE6q1+/FDkHpxqp1V3so5eE/tIsiyM2+4UjyjhWmzVqn1hI7ZciR2by/3zZR2MvkWvarRuoDh8+XC+99FLr6zlz5qiqqkqFhYXt+losFkUikcu2d0VjY5PC4fbjxJLTmaGGhgtxrdGXsJ5d4wta5fX6O+yTnp4WtU80oVC4x2N0ls8XUEODmVrdwT76lfRQQBebe/6zcjjS1BxlnIxQOGqfWIj4AvIm+c+XfTT24r2mVqulw4NF3QpeH3/8serr61VQUCBJikQistvtysrK0tmzZ1v7eTweuVyudu0NDQ2tpy2B3iwkq/yGvgcS318rAAC9QbeCVyQS0XPPPafx48fL4XBoy5YtuvfeezVmzBgdP35cJ06c0LXXXqsdO3ZoxowZGjZsmNLS0nTo0CHdfvvtqqqqUm5ubqy3BYg5f1CqOXA2escYuGfs1UbqAAASp1vBa+TIkXrwwQc1a9YsBYNB5efna8qUKZKk1atXa/78+fL7/crLy2s9/VhRUaElS5bI6/Vq9OjRKi0tjd1WAAAAJIEuBa+9e/e2/rmkpEQlJSXt+kyYMEHbt29v1z5y5Eht3bq1G1MEAAC4MnDnegAAAEO6daoRQPLKTPEp93oz3xgYaPMpqN59OwkAMIngBfQxtqBPje+buYHxsCnTFLQTvADgS5xqBAAAMITgBQAAYAinGgGgl+pv8csa9BmpZbOEjNQB+jqCFwD0UtagTxd/v8tIrYwxdxmpA/R1nGoEAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMCQTgWvpqYmTZkyRSdPnpQk1dbWqri4WPn5+Vq3bl1rv6NHj2rGjBkqKCjQ4sWLFQwGJUmnT59WSUmJCgsL9fDDD8vr9cZhUwAAAHq3qMHr8OHDmjVrlurr6yVJPp9P5eXl2rhxo6qrq1VXV6d9+/ZJksrKyrR06VLt3LlTkUhElZWVkqQVK1Zo9uzZqqmp0U033aSNGzfGb4sAAAB6qajBq7KyUsuWLZPL5ZIkHTlyRNnZ2Ro+fLjsdruKi4tVU1OjU6dOyefzKScnR5I0ffp01dTUKBAI6MMPP1RBQVOvQFEAAA0nSURBVEGbdgAAgL7GHq3DqlWr2rz2eDxyOp2tr10ul9xud7t2p9Mpt9utc+fOacCAAbLb7W3au2rIkAFd/kx3OJ0ZRur0Fcm+nu5zLUpPTzNSy2azdqpWT+djsYSUkhL1f/2YsNmsvX4f6M3zC/2lWRaHmf3ParPKEaNa0caJZa2OpPZLkSOz9/58O6s376PJKpFr2uW/fSORSLs2i8XS5fauamxsUjjcfqxYcjoz1NBwIa41+pIrYT19Qau8Xr+RWqFQOGqt9PS0Hs8nErErEAj2aIzOCoXCvXof6O37aHoooIvNZva/jFBYzTGo5XCkRR0nVrWiifgC8vbin29n9PZ9NBnFe02tVkuHB4u6/K3GrKwsnT17tvW1x+ORy+Vq197Q0CCXy6WrrrpKTU1NCoVCbdoBAAD6mi4HrzFjxuj48eM6ceKEQqGQduzYodzcXA0bNkxpaWk6dOiQJKmqqkq5ublKSUnR2LFjVV1d3aYdAACgr+nyqca0tDStXr1a8+fPl9/vV15engoLCyVJFRUVWrJkibxer0aPHq3S0lJJ0rJly7Ro0SJt2rRJQ4cO1dq1a2O7FQAAAEmg08Fr7969rX+eMGGCtm/f3q7PyJEjtXXr1nbtw4YN0+bNm7s5RQAAgCsDd64HAAAwhOAFAABgCMELAADAEIIXAACAIQQvAAAAQwheAAAAhhC8AAAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIYQvAAAAAwheAEAABhC8AIAADCE4AUAAGAIwQsAAMAQghcAAIAh9kRPAACAeLDZLEoPnTdSK2zvp4uRNCO1kNwIXgCAK5Il6NfFw+8YqdU/J1+yEbwQHacaAQAADCF4AQAAGELwAgAAMITgBQAAYEiPLq4vLS1VY2Oj7PZLw6xcuVKfffaZNm3apEAgoLlz56qkpESSVFtbq+eff15+v1+TJk3SwoULez57AEiA/ha/rEFf3OvYLKG41wBgVreDVyQS0aeffqp33nmnNXi53W4tXLhQv/71r5WamqqZM2fqzjvv1LXXXqvy8nJt3rxZQ4cO1bx587Rv3z7l5eXFbEMAwBRr0KeLv98V9zoZY+6Kew0AZnU7eH366aeyWCx64IEH1NjYqO9973tKT0/X+PHjlZmZKUkqKChQTU2Nxo0bp+zsbA0fPlySVFxcrJqaGoIXAADoU7odvL744gtNmDBBy5cvl8/nU2lpqSZNmiSn09nax+Vy6ciRI/J4PO3a3W53l+oNGTKgu1PtEqczw0idviLZ19N9rkXp6WbuzWOzWTtVq6fzsVhCSkkxcws/m83a6/eB7swv9JdmWRzx3y+sNqscBurEula0cUxtl8n1S+2XIkdmfPb13v7/UDJK5Jp2+2/fW2+9VbfeeqskyeFw6L777tPzzz+vhx56qE0/i8WiSCTS7vMWi6VL9RobmxQOtx8nlpzODDU0XIhrjb7kSlhPX9Aqr9dvpFYoFI5aKz09rcfziUTsCgSCPRqjs0KhcK/eB7q7j6aHArrYHP/9IiMUVrOBOrGs5XCkRR3H1HaZXL+ILyBvHPb1K+Hv0d4m3mtqtVo6PFjU7W81Hjx4UPv37299HYlENGzYMJ09e7a1zePxyOVyKSsr67LtAAAAfUm3g9eFCxe0Zs0a+f1+NTU16c0339Q///M/a//+/fr888918eJF7dq1S7m5uRozZoyOHz+uEydOKBQKaceOHcrNzY3ldgAAAPR63T7VePfdd+vw4cOaNm2awuGwZs+erdtvv10LFy5UaWmpAoGA7rvvPt1yyy2SpNWrV2v+/Pny+/3Ky8tTYWFhzDYCAAAgGfToCtsf/vCH+uEPf9imrbi4WMXFxe36TpgwQdu3b+9JOeCKlpniU+71HV97ZbdHFAz27N5OqdauXV8JAIgdM19tAhCVLehT4/s1HfZJSen5hfHXFU3u0ecBAN3HI4MAAAAMIXgBAAAYQvACAAAwhGu8AFwRuvPg6tBfmpUeCnS5Fg+vBtBdBC8AV4TuPLja4kjr1h3oeXg1gO7iVCMAAIAhBC8AAABDCF4AAACGELwAAAAM4eJ6AHGTYpPsofNGavFNQwDJgOAFIG4soRY1H/6tkVp80xBAMiB4AQDQQzabRelxOLp7uXvNhe39dDGSFvNaMIPgBQBAD1mCfl08/E7sx73Mveb65+RLNoJXsiJ4AQCQROJ1dO2bOLIWHwQvAACSSLyOrn0TR9big9tJAAAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCF8qxEAALRj6rYVUt+6dQXBCwAAtGPqthVS37p1BcELQFwFwhYjdSIyUwcAeoLgBXRgoM2n3OuDRmqlWq+84BCJSH881Wyk1thbjZQBgB4heCHphGSV30wWUqZ8any/xkit64omG6kDAEgco8Hrrbfe0qZNmxQIBDR37lyVlJSYLI8rhD8o1Rw4a6TW/76D302QOBFZOFWLPsHohfwXjZT5q4z9q+J2u7Vu3Tr9+te/VmpqqmbOnKk777xTI0aMMDUFAEgqEXGqFn2DyQv508YXSXIYqXU5xoJXbW2txo8fr8zMTElSQUGBampq9Oijj3bq81ZD17+YqtNXxGM9bTaLBvS3xXzcy7HabEobkGGklsVmj1rLnmKTNRCKe51YMV3L2n9A1z6TliprJKXLtSK2FIXT4r9dptcvFtt0UVaF01I77BOxpXT5Z9Ut3dgneluty+6jprbrCli/y7Ja4/pvfbSxLZFIJBK36l/z05/+VM3NzVq4cKEk6T//8z915MgRPfPMMybKAwAAJJyxO9dfLt9ZLBxdAgAAfYex4JWVlaWzZ7+6INrj8cjlcpkqDwAAkHDGgtd3vvMd7d+/X59//rkuXryoXbt2KTc311R5AACAhDN2cX1WVpYWLlyo0tJSBQIB3XfffbrllltMlQcAAEg4YxfXAwAA9HXGTjUCAAD0dQQvAAAAQwheAAAAhhC8AAAADCF4fc3Bgwc1ffp0FRcX66GHHtL582Ye2HmlOnTokGbMmKGpU6fq+9//vk6dOpXoKV0x1q9fr3/9139N9DSS2ltvvaXJkydr4sSJeuONNxI9nStCU1OTpkyZopMnTyZ6KleEF198UUVFRSoqKtKaNWsSPZ2kt379ek2ePFlFRUV69dVXEzYPgtfXPPXUU1qzZo3eeustjRgxQi+//HKip5TUysrKtGrVKm3btk3FxcV69tlnEz2lpHfhwgWVl5frlVdeSfRUkprb7da6dev0i1/8Qtu2bdOWLVv0xz/+MdHTSmqHDx/WrFmzVF9fn+ipXBFqa2v17rvv6s0331RVVZU++ugj7d69O9HTSloHDhzQ+++/r+3bt+tXv/qVNm/erE8//TQhcyF4fU11dbVGjBihQCAgt9utgQMHJnpKSaulpUWPPfaYRo4cKUn61re+pTNnziR4Vslvz549uu666/QP//APiZ5KUqutrdX48eOVmZkph8OhgoIC1dTUJHpaSa2yslLLli3jiSQx4nQ6tWjRIqWmpiolJUU33HCDTp8+nehpJa1x48bptddek91uV2Njo0KhkBwOR0LmQvD6mpSUFH388cfKy8vTBx98oKKiokRPKWmlpqZq6tSpkqRwOKwXX3xR99xzT4JnlfymTZumBx98UDabLdFTSWoej0dOp7P1tcvlktvtTuCMkt+qVas0duzYRE/jinHjjTcqJydHklRfX6/q6mrl5eUleFbJLSUlRRs2bFBRUZEmTJigrKyshMyjTwav3/zmN8rNzW3z39y5cyVdOjJTW1urf/zHf9TChQsTO9Ek0dF6trS06Ec/+pGCwaDmzZuX2IkmkY7WFD13uftGWyyWBMwE6Ngnn3yiH/zgB3ryySd13XXXJXo6SW/BggXav3+/zpw5o8rKyoTMwdgjg3qTSZMmadKkSW3a/H6/3n777dajMt/97nf1wgsvJGJ6Sedy6ylJXq9XDz/8sDIzM7Vp0yalpKQkYHbJ6a+tKWIjKytLBw8ebH3t8Xg4RYZe59ChQ1qwYIHKy8s5A9NDx44dU0tLi0aNGqX+/fsrPz9fH3/8cULm0iePeF2O3W7XihUrVFdXJ+nSEYfbbrstwbNKbmVlZcrOztb69euVmpqa6OkArb7zne9o//79+vzzz3Xx4kXt2rVLubm5iZ4W0OrMmTN65JFHVFFRQeiKgZMnT2rJkiVqaWlRS0uL9uzZo9tvvz0hc+mTR7wux2azad26dXr66acVCoWUlZWlVatWJXpaSesPf/iD9uzZoxEjRmjatGmSLl1H87Of/SzBMwMuHfFauHChSktLFQgEdN999+mWW25J9LSAVi+//LL8fr9Wr17d2jZz5kzNmjUrgbNKXnl5eTp8+LCmTZsmm82m/Pz8hAVaHpINAABgCKcaAQAADCF4AQAAGELwAgAAMITgBQAAYAjBCwAAwBCCFwAAgCEELwAAAEMIXgAAAIb8/7BZvtuMzlnQAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "for col in pc1_features:\n", " bins = np.linspace(-3, 3, 20)\n", " fig, ax = plt.subplots(1,1, figsize=(10,5))\n", " X_std.loc[left_group, col].hist(ax=ax, bins=bins, color='C0', alpha=0.5, label='Group 1')\n", " X_std.loc[~left_group, col].hist(ax=ax, bins=bins, color='C1', alpha=0.5, label='Group 2')\n", " plt.legend(loc='best')\n", " plt.title(col)\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the most significant difference is between the longitute and latitude. Another interesting observation is that in the first group (blue) the rooms per household seem to be slightly lower and the bedrooms per room slightly higher. We will discuss further below. First we want to look at the longitude and latitude distribution of the two groups. First we make a hexplot to see the hotspots." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Group 1\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Group 2\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.gridspec as gridspec\n", "\n", "print(\"Group 1\")\n", "sns.jointplot(\n", " kind='hex',\n", " x=\"longitude\",\n", " y=\"latitude\",\n", " cmap='magma',\n", "\n", " data=X.loc[left_group],\n", ")\n", "plt.show()\n", "\n", "print(\"Group 2\")\n", "sns.jointplot(\n", " kind='hex',\n", " x=\"longitude\",\n", " y=\"latitude\",\n", " cmap='magma',\n", " data=X.loc[~left_group],\n", ")\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We see that in the first group there are mainly two hotspots while in the second group the hotspots seem to be further distributed. Recall that we can plot the longitude and latitude on top of a map. Let's investigate what these hot regions correspond to." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# ref - www.kaggle.com/camnugent/geospatial-feature-engineering-and-visualization/\n", "california_img=mpimg.imread('images/california.png')\n", "fig, axes = plt.subplots(1, 2, figsize=(12,6))\n", "\n", "fig = sns.scatterplot(\n", " x=\"longitude\",\n", " y=\"latitude\",\n", " data=X.loc[left_group],\n", " alpha=0.1,\n", " s=1,\n", " color=\"r\",\n", " edgecolor=\"none\", \n", " ax=axes[0]\n", ")\n", "\n", "axes[0].imshow(california_img, extent=[-124.55, -113.80, 32.45, 42.05], alpha=0.5)\n", "axes[0].set_title(\"Group 1\")\n", "fig = sns.scatterplot(\n", " x=\"longitude\",\n", " y=\"latitude\",\n", " data=X.loc[~left_group],\n", " alpha=0.05,\n", " s=3,\n", " color=\"r\",\n", " edgecolor=\"none\",\n", " ax=axes[1]\n", "\n", ")\n", "axes[1].imshow(california_img, extent=[-124.55, -113.80, 32.45, 42.05], alpha=0.5)\n", "axes[1].set_title(\"Group 2\")\n", "\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Comparing to a satellite image of California (see below) we see that the two hotspots in the first group correspond to Los Angeles and San Diego. The second group mainly includes San Francisco, the Silicon Valley, Sacramento, Fresno and Bakersfield. What is interesting to observe is that the second group also includes Riverside which is just next to Los Angeles. That means that the second principal component is not just the diagonal along the coast but specifically encodes Los Angeles and San Diege. If one wanted to build a classifier to determine whether a house is in Los Angeles or an other region of California one would only need to look at the second principle component of the dataset.\n", "\n", "Circling back the observation about the rooms per households. It seems that in Los Angeles the appartments have slightly fewer rooms than the rest of California but a similar amount of bedrooms, hence the higher number of bedrooms per rooms. In other words in Los Angeles you would get less extra rooms per household.\n", "\n", "
\n", "\n", "

Figure: Satellite image of California from Google Maps.

\n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Finally, we can also look at the median value of the houses in the two groups:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "bins=50\n", "fig, ax = plt.subplots(1,1, figsize=(10,5))\n", "y.loc[left_group].hist(ax=ax, bins=bins, color='C0', alpha=0.5, label='Group 1')\n", "y.loc[~left_group].hist(ax=ax, bins=bins, color='C1', alpha=0.5, label='Group 2')\n", "plt.legend(loc='best')\n", "plt.title('median value')\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "One observes that the median value is higher for the first group (Los Angeles & San Diego). However, the long tail in both groups is quite similar. So living seems to be more expensive for low and medium income people while for the upper class the prices seem to be comparable." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Compressed model\n", "Now we want to use PCA to build a model that is more compact than the model that utilizes all features. We add one principal component after the other and measure how well a Random Forest performs with this subset of features. We do this in our standard cross-validation loop." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 17/17 [04:18<00:00, 15.18s/it]\n" ] } ], "source": [ "rmse_valid =[]\n", "for i in tqdm(range(1, 18)):\n", " rf = RandomForestRegressor(n_estimators=100, max_features='sqrt', n_jobs=-1, random_state=42)\n", " \n", " results = cross_validate(rf, X_pca[:, :i], y, cv=5,\n", " return_train_score=True,\n", " scoring='neg_root_mean_squared_error')\n", " rmse_valid.append(-np.mean(results[\"test_score\"]))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As an additional baseline we also create a median regressor:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "rmse_median = rmse(y, [np.median(y)]*len(y))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we can plot the performances with the full Random Forest and the median regresssor as baselines for all subsets of components." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "lineplot = sns.lineplot(x=np.arange(len(rmse_valid))+1 , y=rmse_valid, label='pca')\n", "lineplot.axhline(rmse_median, c='y', linestyle='--', label='median')\n", "lineplot.axhline(rmse_full, c='r', linestyle='--', label='full model')\n", "plt.legend(loc='best')\n", "plt.xlabel('number of PCA components')\n", "plt.ylabel('RMSE');" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "So we can see that already with 4-7 principal components we get results that are quite close to the full model. As mentioned previously this has two advantages:\n", "1. Visualising 4 components is still hard but feasable compared to 17\n", "2. Training a Random Forest with half the features is twice as fast for training and inference.\n", "\n", "Looking at the graph we see that there are several steep steps. It seems that adding these components improved the performance significantly. By taking the components with the biggest steps we can try to build a better selection of features:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "features = [0,1,3,6,10]" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "rf = RandomForestRegressor(n_estimators=100, max_features='sqrt', n_jobs=-1, random_state=42)\n", " \n", "results = cross_validate(rf, X_pca[:, features], y, cv=5,\n", " return_train_score=True,\n", " scoring='neg_root_mean_squared_error')\n", "rmse_pca = -np.mean(results[\"test_score\"])" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "64700.242256282574" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "rmse_pca" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Just with 5 of 17 features we get <10% of the model that used all features. This model is 3-4 times faster to train and to make predictions. In practice there are more systematic ways of finding the best subset of features such as *forward* and *backward selection* (see [link](https://www.kdnuggets.com/2018/06/step-forward-feature-selection-python.html))." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Exercise 2\n", "Make your own selection of principle components (e.g. the first five) and see if you can beat the selection above." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" } }, "nbformat": 4, "nbformat_minor": 4 }