2-1 Regression.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Homework 2-1.ipynb","provenance":[],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"}},"cells":[{"cell_type":"markdown","metadata":{"id":"mwNEDkhFtxwC"},"source":["# <font style=\"font-size:40px\"><center><b>Machine Learning</b></center></font><font                        style=\"font-size:30px\"><center><b>Master in Computational Finance (MCF) </b></center></font> \n","\n","<center> \n","\n","# ![Screen Shot 2021-07-11 at 13.37.02 PM.png](data:image/png;base64,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)\n","\n","<center/>"]},{"cell_type":"markdown","metadata":{"id":"aihhzeTmt6TX"},"source":["# <font color='tomato' style=\"font-size:40px\"><center><b>Homework 2</b></center></font></b></center></font> \n","\n","<center>\n","<b>profesor: </b> Nikola Vasiljevic\n","\n","<b>students: </b> \n","Nikola Krivacevic, Aleksandar Milinkovic\n","</center>\n","\n","\n","\n"]},{"cell_type":"markdown","metadata":{"id":"DCLM0li4vwQO"},"source":["## Importing libraries:\n"]},{"cell_type":"code","metadata":{"id":"GtDR2Pko4-Tf"},"source":["import warnings\n","warnings.filterwarnings('ignore')"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"vz8fhU86vu5w"},"source":["import numpy as np\n","import pandas as pd  \n","from scipy.special import comb\n","import datetime as dt\n","import statsmodels.api as sm\n","\n","from sklearn.pipeline import Pipeline\n","from sklearn.pipeline import make_pipeline\n","\n","from sklearn.preprocessing import StandardScaler\n","from sklearn.preprocessing import PolynomialFeatures\n","\n","from sklearn.feature_selection import f_regression\n","from sklearn.feature_selection import mutual_info_regression\n","from sklearn.feature_selection import (SelectKBest, SelectPercentile, RFE, RFECV)\n","\n","from sklearn.linear_model import (LinearRegression, Ridge, RidgeCV, Lasso, LassoCV, ElasticNet, ElasticNetCV)\n","from sklearn.cross_decomposition import PLSRegression\n","\n","from sklearn.model_selection import cross_val_score\n","from sklearn.model_selection import cross_validate\n","from sklearn.metrics import mean_squared_error\n","\n","from sklearn import metrics\n","\n","import matplotlib.pyplot as plt\n","import seaborn as sns\n","import plotly.graph_objects as go\n","from plotly.subplots import make_subplots"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["## Supervised Learning"],"metadata":{"id":"uA4MDZwE8Xa_"}},{"cell_type":"markdown","source":["### Consider the California Housing Dataset that was studied and extensively discused in the lectures."],"metadata":{"id":"Wh92dpE88bze"}},{"cell_type":"markdown","source":["### (a) Data preprocessing: Provide a proposal on how to improve data preprocessing relative to the procedure described in the lecture notes. Evaluate the impact of different data treatment algorithms on the OLS regression results using the full dataset.Leverage the code provided in lecture notes as much as you can/want. Motivate your choices and discuss the results."],"metadata":{"id":"j9pPOh5m8hkD"}},{"cell_type":"markdown","source":["#### California housing dataset EDA"],"metadata":{"id":"BPT24KMeSR7B"}},{"cell_type":"markdown","source":["We'll work on California housing dataset and will try to model the linear relationship between house properities and house price. By understanding the semantics of the features we have in dataset, we can conclude that the observations are not single houses, but median values for California districts/counties."],"metadata":{"id":"YKTehkPAeINI"}},{"cell_type":"markdown","source":["Our `target variable` is house price - `MedHouseVal`.\n","\n","Dataset `features/regressors`:\n","- MedInc - median income for the district\n","- House age\n","- AveRooms - average num. of rooms in the house\n","- Population - number of people living in the district\n","- AveOccup - average number of tenants in the house\n","- Latitude/Logitude - geo location of county"],"metadata":{"id":"i3W5ZqV6e5vm"}},{"cell_type":"markdown","source":["##### Dataset statistical properties"],"metadata":{"id":"AZY4yS4XJ_kW"}},{"cell_type":"code","source":["# California housing dataset\n","from sklearn.datasets import fetch_california_housing\n","california_housing = fetch_california_housing()"],"metadata":{"id":"nhTCFVcI9OHV"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Create a dataframe containing observations\n","data = pd.DataFrame(california_housing.data, columns=california_housing.feature_names)\n","data['MedHouseVal'] = california_housing.target\n","data.describe()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":300},"id":"-Tz5Qlyq9OSv","executionInfo":{"status":"ok","timestamp":1639534105594,"user_tz":-60,"elapsed":8,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"6d0d126b-fdbf-4916-9d59-470c22cb3de3"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>count</th>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","      <td>20640.000000</td>\n","    </tr>\n","    <tr>\n","      <th>mean</th>\n","      <td>3.870671</td>\n","      <td>28.639486</td>\n","      <td>5.429000</td>\n","      <td>1.096675</td>\n","      <td>1425.476744</td>\n","      <td>3.070655</td>\n","      <td>35.631861</td>\n","      <td>-119.569704</td>\n","      <td>2.068558</td>\n","    </tr>\n","    <tr>\n","      <th>std</th>\n","      <td>1.899822</td>\n","      <td>12.585558</td>\n","      <td>2.474173</td>\n","      <td>0.473911</td>\n","      <td>1132.462122</td>\n","      <td>10.386050</td>\n","      <td>2.135952</td>\n","      <td>2.003532</td>\n","      <td>1.153956</td>\n","    </tr>\n","    <tr>\n","      <th>min</th>\n","      <td>0.499900</td>\n","      <td>1.000000</td>\n","      <td>0.846154</td>\n","      <td>0.333333</td>\n","      <td>3.000000</td>\n","      <td>0.692308</td>\n","      <td>32.540000</td>\n","      <td>-124.350000</td>\n","      <td>0.149990</td>\n","    </tr>\n","    <tr>\n","      <th>25%</th>\n","      <td>2.563400</td>\n","      <td>18.000000</td>\n","      <td>4.440716</td>\n","      <td>1.006079</td>\n","      <td>787.000000</td>\n","      <td>2.429741</td>\n","      <td>33.930000</td>\n","      <td>-121.800000</td>\n","      <td>1.196000</td>\n","    </tr>\n","    <tr>\n","      <th>50%</th>\n","      <td>3.534800</td>\n","      <td>29.000000</td>\n","      <td>5.229129</td>\n","      <td>1.048780</td>\n","      <td>1166.000000</td>\n","      <td>2.818116</td>\n","      <td>34.260000</td>\n","      <td>-118.490000</td>\n","      <td>1.797000</td>\n","    </tr>\n","    <tr>\n","      <th>75%</th>\n","      <td>4.743250</td>\n","      <td>37.000000</td>\n","      <td>6.052381</td>\n","      <td>1.099526</td>\n","      <td>1725.000000</td>\n","      <td>3.282261</td>\n","      <td>37.710000</td>\n","      <td>-118.010000</td>\n","      <td>2.647250</td>\n","    </tr>\n","    <tr>\n","      <th>max</th>\n","      <td>15.000100</td>\n","      <td>52.000000</td>\n","      <td>141.909091</td>\n","      <td>34.066667</td>\n","      <td>35682.000000</td>\n","      <td>1243.333333</td>\n","      <td>41.950000</td>\n","      <td>-114.310000</td>\n","      <td>5.000010</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["             MedInc      HouseAge  ...     Longitude   MedHouseVal\n","count  20640.000000  20640.000000  ...  20640.000000  20640.000000\n","mean       3.870671     28.639486  ...   -119.569704      2.068558\n","std        1.899822     12.585558  ...      2.003532      1.153956\n","min        0.499900      1.000000  ...   -124.350000      0.149990\n","25%        2.563400     18.000000  ...   -121.800000      1.196000\n","50%        3.534800     29.000000  ...   -118.490000      1.797000\n","75%        4.743250     37.000000  ...   -118.010000      2.647250\n","max       15.000100     52.000000  ...   -114.310000      5.000010\n","\n","[8 rows x 9 columns]"]},"metadata":{},"execution_count":4}]},{"cell_type":"code","source":["data.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"JR5bfSoo9OcE","executionInfo":{"status":"ok","timestamp":1639534105595,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"85d3bada-3dad-4501-de98-6f248306371e"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>8.3252</td>\n","      <td>41.0</td>\n","      <td>6.984127</td>\n","      <td>1.023810</td>\n","      <td>322.0</td>\n","      <td>2.555556</td>\n","      <td>37.88</td>\n","      <td>-122.23</td>\n","      <td>4.526</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>8.3014</td>\n","      <td>21.0</td>\n","      <td>6.238137</td>\n","      <td>0.971880</td>\n","      <td>2401.0</td>\n","      <td>2.109842</td>\n","      <td>37.86</td>\n","      <td>-122.22</td>\n","      <td>3.585</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>7.2574</td>\n","      <td>52.0</td>\n","      <td>8.288136</td>\n","      <td>1.073446</td>\n","      <td>496.0</td>\n","      <td>2.802260</td>\n","      <td>37.85</td>\n","      <td>-122.24</td>\n","      <td>3.521</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>5.6431</td>\n","      <td>52.0</td>\n","      <td>5.817352</td>\n","      <td>1.073059</td>\n","      <td>558.0</td>\n","      <td>2.547945</td>\n","      <td>37.85</td>\n","      <td>-122.25</td>\n","      <td>3.413</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>3.8462</td>\n","      <td>52.0</td>\n","      <td>6.281853</td>\n","      <td>1.081081</td>\n","      <td>565.0</td>\n","      <td>2.181467</td>\n","      <td>37.85</td>\n","      <td>-122.25</td>\n","      <td>3.422</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["   MedInc  HouseAge  AveRooms  ...  Latitude  Longitude  MedHouseVal\n","0  8.3252      41.0  6.984127  ...     37.88    -122.23        4.526\n","1  8.3014      21.0  6.238137  ...     37.86    -122.22        3.585\n","2  7.2574      52.0  8.288136  ...     37.85    -122.24        3.521\n","3  5.6431      52.0  5.817352  ...     37.85    -122.25        3.413\n","4  3.8462      52.0  6.281853  ...     37.85    -122.25        3.422\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":5}]},{"cell_type":"code","source":["data.hist(figsize=(16, 12), bins=40, edgecolor='black')\n","plt.subplots_adjust(hspace=0.3, wspace=0.2)\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":716},"id":"RUD586xCc8zW","executionInfo":{"status":"ok","timestamp":1639534108670,"user_tz":-60,"elapsed":3081,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"c31a43a8-7541-4300-ea14-7a26b9fef659"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 1152x864 with 9 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["We can see that in our dataset observations we have some data that could potentially seen as outliers. We could try to remove them and see if our OLS regression model would be improved compared the the OLS regression model fitter on original, non-transformed dataset."],"metadata":{"id":"NZymH12udbVs"}},{"cell_type":"code","source":["corr = data.drop(columns=[\"MedHouseVal\"]).corr()\n","\n","# Generate a mask for the upper triangle\n","mask = np.triu(np.ones_like(corr, dtype=bool))\n","\n","# Set up the matplotlib figure\n","f, ax = plt.subplots(figsize=(11, 9))\n","\n","# Generate a custom diverging colormap\n","cmap = sns.diverging_palette(230, 20, as_cmap=True)\n","\n","# Draw the heatmap with the mask and correct aspect ratio\n","sns.heatmap(corr, mask=mask, cmap=cmap, vmax=.3, center=0,\n","            square=True, linewidths=.5, cbar_kws={\"shrink\": .5})"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":554},"id":"c7l56budd6GS","executionInfo":{"status":"ok","timestamp":1639534108670,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"5e8c7d2c-33cb-411f-8a09-6110639e1feb"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7fb985329850>"]},"metadata":{},"execution_count":7},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 792x648 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["From the correlation above we can see that the correalation between features in mostly low, around zero.\n","\n","Exception to that is latitude and longitude which is highly negatively correlated (due to geographical direcation of California state).\n","\n","Maybe to explore some reasoning about some non-zero correlations in our dataset:\n","- Low positive corelation between `AvgBedrooms` amd `MedInc` - that would makse sense as it's expected that individials and families who have a higher income to be living in a bigger housed\n","- `AvgBedrooms` and `AvgRooms`, although we don't know for sure how that are different with what they represent, we can see for that pair of features that they also have low, non-negative corerlation."],"metadata":{"id":"LjD0DEM4QygQ"}},{"cell_type":"code","source":["plt.figure(figsize=(20, 40))\n","\n","for i, column in enumerate(data.columns):\n","    # Create subplots\n","    plt.subplot(11, 2, i + 1)\n","    data[column].plot.box(vert=False)\n","    # Labels and graphics\n","    plt.xlabel(None)\n","    plt.ylabel('Box plot of features')\n","    plt.title(column)\n","    plt.grid(color='white', linestyle='-.', linewidth=0.5)\n","    ax = plt.gca()\n","    ax.set_facecolor('lightgrey')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"id":"vgte_jFrtMch","executionInfo":{"status":"ok","timestamp":1639534112342,"user_tz":-60,"elapsed":3677,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"819cf75f-e025-4b20-ffd1-1c57e2c61e45"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 1440x2880 with 9 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["From the Box plots of features, we can see that there could be some outliers (points outside of box plot \"whiskers\"; observations are considered outliers if their distance from the 25th/75th percentile if greater than 1.5 * IRQ)."],"metadata":{"id":"emg7kr2JYlOC"}},{"cell_type":"markdown","source":["##### Removing outliers"],"metadata":{"id":"iEhIu-naKEeK"}},{"cell_type":"markdown","source":["Now let's go feature by feature and try to remove the observations that seem to be outliers:"],"metadata":{"id":"dNumLO-5JuD5"}},{"cell_type":"markdown","source":["###### Average occupancy feature"],"metadata":{"id":"O-lH9hQcKIwl"}},{"cell_type":"code","source":["data['AveOccup'].plot()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"TP1RC-_UJnA3","executionInfo":{"status":"ok","timestamp":1639534112343,"user_tz":-60,"elapsed":13,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"0f8c3321-f0f2-4835-8a1b-51db577fa0d4"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7fb97bae1ad0>"]},"metadata":{},"execution_count":9},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# Selecting which observations to remove based on the criteria\n","observations_to_remove = data[data['AveOccup'] > 8]\n","data_cleaned = data.drop(observations_to_remove.index)\n","print(f'Removing {len(observations_to_remove)} observations.')\n","observations_to_remove.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"gu83Jm74KMXF","executionInfo":{"status":"ok","timestamp":1639534112343,"user_tz":-60,"elapsed":5,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"28e39e54-0ba2-49b3-de2a-f1e123568991"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Removing 53 observations.\n"]},{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>91</th>\n","      <td>0.8668</td>\n","      <td>52.0</td>\n","      <td>2.443182</td>\n","      <td>0.988636</td>\n","      <td>904.0</td>\n","      <td>10.272727</td>\n","      <td>37.80</td>\n","      <td>-122.28</td>\n","      <td>1.375</td>\n","    </tr>\n","    <tr>\n","      <th>270</th>\n","      <td>3.8750</td>\n","      <td>33.0</td>\n","      <td>3.021277</td>\n","      <td>0.659574</td>\n","      <td>575.0</td>\n","      <td>12.234043</td>\n","      <td>37.78</td>\n","      <td>-122.18</td>\n","      <td>2.250</td>\n","    </tr>\n","    <tr>\n","      <th>460</th>\n","      <td>1.4012</td>\n","      <td>52.0</td>\n","      <td>3.105714</td>\n","      <td>1.060000</td>\n","      <td>3337.0</td>\n","      <td>9.534286</td>\n","      <td>37.87</td>\n","      <td>-122.26</td>\n","      <td>1.750</td>\n","    </tr>\n","    <tr>\n","      <th>1039</th>\n","      <td>4.3029</td>\n","      <td>16.0</td>\n","      <td>6.090592</td>\n","      <td>1.121951</td>\n","      <td>4930.0</td>\n","      <td>17.177700</td>\n","      <td>38.42</td>\n","      <td>-120.97</td>\n","      <td>1.219</td>\n","    </tr>\n","    <tr>\n","      <th>1067</th>\n","      <td>3.4191</td>\n","      <td>52.0</td>\n","      <td>4.530612</td>\n","      <td>0.816327</td>\n","      <td>1107.0</td>\n","      <td>11.295918</td>\n","      <td>39.73</td>\n","      <td>-121.85</td>\n","      <td>1.375</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["      MedInc  HouseAge  AveRooms  ...  Latitude  Longitude  MedHouseVal\n","91    0.8668      52.0  2.443182  ...     37.80    -122.28        1.375\n","270   3.8750      33.0  3.021277  ...     37.78    -122.18        2.250\n","460   1.4012      52.0  3.105714  ...     37.87    -122.26        1.750\n","1039  4.3029      16.0  6.090592  ...     38.42    -120.97        1.219\n","1067  3.4191      52.0  4.530612  ...     39.73    -121.85        1.375\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":10}]},{"cell_type":"markdown","source":["###### Average rooms feature"],"metadata":{"id":"fxMKlvaeLl2j"}},{"cell_type":"code","source":["data['AveRooms'].plot()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"lP_7rCD8YGro","executionInfo":{"status":"ok","timestamp":1639534112970,"user_tz":-60,"elapsed":631,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"82d782ea-f5dd-43ae-b203-ae10172c9b58"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7fb97a021210>"]},"metadata":{},"execution_count":11},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# Selecting which observations to remove based on the criteria\n","observations_to_remove = data_cleaned[data_cleaned['AveRooms'] > 10]\n","data_cleaned = data_cleaned.drop(observations_to_remove.index)\n","print(f'Removing {len(observations_to_remove)} observations.')\n","observations_to_remove.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"rsVuvbaoLDa_","executionInfo":{"status":"ok","timestamp":1639534112970,"user_tz":-60,"elapsed":5,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"8f51c1e0-2d1d-4912-a0d8-8231ac3f230a"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Removing 224 observations.\n"]},{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>1023</th>\n","      <td>2.2417</td>\n","      <td>15.0</td>\n","      <td>10.515306</td>\n","      <td>2.372449</td>\n","      <td>573.0</td>\n","      <td>2.923469</td>\n","      <td>38.72</td>\n","      <td>-119.93</td>\n","      <td>0.979</td>\n","    </tr>\n","    <tr>\n","      <th>1024</th>\n","      <td>3.1500</td>\n","      <td>16.0</td>\n","      <td>29.852941</td>\n","      <td>5.323529</td>\n","      <td>202.0</td>\n","      <td>1.980392</td>\n","      <td>38.52</td>\n","      <td>-120.00</td>\n","      <td>1.406</td>\n","    </tr>\n","    <tr>\n","      <th>1030</th>\n","      <td>3.0125</td>\n","      <td>15.0</td>\n","      <td>10.168591</td>\n","      <td>2.057737</td>\n","      <td>1103.0</td>\n","      <td>2.547344</td>\n","      <td>38.55</td>\n","      <td>-120.25</td>\n","      <td>1.117</td>\n","    </tr>\n","    <tr>\n","      <th>1102</th>\n","      <td>2.4028</td>\n","      <td>17.0</td>\n","      <td>31.777778</td>\n","      <td>9.703704</td>\n","      <td>47.0</td>\n","      <td>1.740741</td>\n","      <td>40.06</td>\n","      <td>-121.54</td>\n","      <td>0.675</td>\n","    </tr>\n","    <tr>\n","      <th>1233</th>\n","      <td>3.7500</td>\n","      <td>10.0</td>\n","      <td>16.847534</td>\n","      <td>3.237668</td>\n","      <td>546.0</td>\n","      <td>2.448430</td>\n","      <td>38.23</td>\n","      <td>-120.34</td>\n","      <td>1.214</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["      MedInc  HouseAge   AveRooms  ...  Latitude  Longitude  MedHouseVal\n","1023  2.2417      15.0  10.515306  ...     38.72    -119.93        0.979\n","1024  3.1500      16.0  29.852941  ...     38.52    -120.00        1.406\n","1030  3.0125      15.0  10.168591  ...     38.55    -120.25        1.117\n","1102  2.4028      17.0  31.777778  ...     40.06    -121.54        0.675\n","1233  3.7500      10.0  16.847534  ...     38.23    -120.34        1.214\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":12}]},{"cell_type":"markdown","source":["###### Average bedrooms feature"],"metadata":{"id":"MLCYicYvM73O"}},{"cell_type":"code","source":["data['AveBedrms'].plot()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"oR8qmQpPYG2A","executionInfo":{"status":"ok","timestamp":1639534113486,"user_tz":-60,"elapsed":519,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"144f493b-7779-429f-a961-522acc056014"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7fb978f872d0>"]},"metadata":{},"execution_count":13},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXAAAAD4CAYAAAD1jb0+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXhU5d038O8vCYsgOxFiRaKARXxVwJSqqHWXwvN0eX3aq3azT+1LbfGpWttLXEtttbR16UOrKCqKK6JAobIICghUFhMMEBKWBAJkHwLZCFlm5vf+MWcmM5OZzMnsJ/P9XFeuzJw5c+aee+Z85z73uc85oqogIiLrSUt0AYiIKDwMcCIii2KAExFZFAOciMiiGOBERBaVEc8XGz58uGZnZ8fzJYmILC8vL++Eqmb6T49rgGdnZyM3NzeeL0lEZHkicjTQ9JBdKCLSV0R2ishuEdknIr83pr8uIkdEJN/4mxjtQhMRUXBmWuCtAG5U1SYR6QVgq4isMR77rap+ELviERFRMCEDXF2HajYZd3sZfzx8k4gowUyNQhGRdBHJB1ADYL2q7jAeelJE9ojIcyLSJ8hzZ4pIrojk2my2KBWbiIhMBbiqOlR1IoDzAEwRkf8D4CEA4wF8BcBQAA8Gee4CVc1R1ZzMzE47UYmIKEzdGgeuqnUANgKYpqqV6tIK4DUAU2JRQCIiCszMKJRMERls3D4LwC0A9otIljFNAHwLQEEsC0pERL7MjELJArBIRNLhCvwlqvqhiGwQkUwAAiAfwN0xLGfcNbS0Y9MBG75x+bmJLgoRUUBmRqHsATApwPQbY1KiJPHAkt1YX1iNi0cOwLgRAxJdHCKiTngulCAq688AAFranQkuCRFRYAxwIiKLYoATEVkUA5yIyKIY4EREFsUAJyKyKAY4EZFFMcCJiCyKAU5EZFEMcCIii2KAExFZFAOciMiiGOBERBbFACcisigGOBGRRTHAiYgsigFORGRRDHAiIotigBMRWRQDnIjIokIGuIj0FZGdIrJbRPaJyO+N6ReIyA4RKRaR90Skd+yLS0REbmZa4K0AblTVywFMBDBNRK4E8GcAz6nqWACnANwVu2ISEZG/kAGuLk3G3V7GnwK4EcAHxvRFAL4VkxISEVFApvrARSRdRPIB1ABYD6AEQJ2q2o1ZygB8KchzZ4pIrojk2my2aJSZiIhgMsBV1aGqEwGcB2AKgPFmX0BVF6hqjqrmZGZmhllMIiLy161RKKpaB2AjgKsADBaRDOOh8wCUR7lsRETUBTOjUDJFZLBx+ywAtwAogivI/8uY7U4AK2JVyERQTXQJiIi6lhF6FmQBWCQi6XAF/hJV/VBECgEsFpE/AvgCwKsxLGfCiCS6BEREgYUMcFXdA2BSgOmH4eoPJyKiBOCRmEREFsUAJyKyKAY4EZFFMcCJiCyKAU5EZFEMcCIii2KAExFZFAOciMiiGOBERBbFACcisigGOBGRRTHAiYgsigFORGRRDHAiIotigBMRWRQDnIjIohjgREQWxQAnIrIoBngQ7osa8+LGRJSsGOABLM0rQ2FlQ6KLQUTUpZABLiKjRGSjiBSKyD4RudeYPkdEykUk3/ibHvvixsfK3RWJLgIRUUghr0oPwA7gAVXdJSIDAOSJyHrjsedU9enYFY+IiIIJGeCqWgmg0rjdKCJFAL4U64IREVHXutUHLiLZACYB2GFMukdE9ojIQhEZEuQ5M0UkV0RybTZbRIUlIqIOpgNcRM4GsBTAfaraAGA+gDEAJsLVQn8m0PNUdYGq5qhqTmZmZhSKTEREgMkAF5FecIX326q6DABUtVpVHarqBPAygCmxK2biiCS6BEREgZkZhSIAXgVQpKrPek3P8prt2wAKol88IiIKxswolKkAfgRgr4jkG9MeBnCHiEwEoABKAfw8JiVMAB67Q0RWYGYUylYAgToSVke/OEREZBaPxCQisigGOBGRRTHAiYgsigFORGRRDHAiIotigAegPAk4EVkAA5yIyKIY4EREFsUAJyKyKAY4EZFFMcCJiCyKAU5EZFEMcCIii2KAExFZFAOciMiiGOBERBbFACcisigGOBGRRTHAiYgsigFORGRRDPAAeDZZIrKCkAEuIqNEZKOIFIrIPhG515g+VETWi8gh4/+Q2BeXiIjczLTA7QAeUNUJAK4EMEtEJgCYDeATVR0H4BPjPhERxUnIAFfVSlXdZdxuBFAE4EsAvglgkTHbIgDfilUhiYios271gYtINoBJAHYAGKGqlcZDVQBGBHnOTBHJFZFcm80WQVGJiMib6QAXkbMBLAVwn6o2eD+mrotIBtz1p6oLVDVHVXMyMzMjKiwREXUwFeAi0guu8H5bVZcZk6tFJMt4PAtATWyKSEREgZgZhSIAXgVQpKrPej20EsCdxu07AayIfvGIiCiYDBPzTAXwIwB7RSTfmPYwgLkAlojIXQCOAvhubIpIRESBhAxwVd0KQII8fFN0i0NERGbxSEwiIotigBMRWRQDnIjIohjgREQWxQAnIrIoBngAGvigUiKipMIAJyKyKAY4EZFFMcCJiCyKAU5EZFEMcCIii2KAExFZFAM8AF6VnoisgAFORGRRDHAiIotigBMRWRQDnIjIohjgREQWxQAnIrIoBjgRkUWFDHARWSgiNSJS4DVtjoiUi0i+8Tc9tsWML44Dp2jZdewU1hZUJboY1EOZaYG/DmBagOnPqepE4291dItF1DP83xc+w91v5SW6GNRDhQxwVd0M4GQcykJERN0QSR/4PSKyx+hiGRJsJhGZKSK5IpJrs9kieDkiIvIWboDPBzAGwEQAlQCeCTajqi5Q1RxVzcnMzAzz5YiIyF9YAa6q1arqUFUngJcBTIlusYiIKJSwAlxEsrzufhtAQbB5iYgoNjJCzSAi7wK4HsBwESkD8DsA14vIRAAKoBTAz2NYxrjjVekpnj7IK8PUscOQNeisRBeFLCZkgKvqHQEmvxqDshClnKZWO37z/m5cMLw/Nv7m+kQXhyyGR2ISJZDTOGrsRGNrgktCVsQAJyKyKAY4URLgXhcKBwOcKIEk0QUgS2OAByBcrYjIAhjgAXAYIcWb8hSYFAYGOFECiXBrj8LHACcisigGOBGRRTHAiZIAe8ApHAxwogRiDzhFggFORGRRDHCiJMBRhBQOBngAXJkoXjiKkCLBAKce7aVPS7D5IK/FSj1TSgW4w6n49KCNR72lkD+t2Y8fL9yZ6GKExKN/KRwpFeAvbzmMOxfuxCdFNYkuChEAnneHIpNSAX60thkAUN3YkuCSEBFFLqUCPBzcyUREyYoBTpQEuFuGwhEywEVkoYjUiEiB17ShIrJeRA4Z/4fEtpjRYm4t4bpE8cItPIqEmRb46wCm+U2bDeATVR0H4BPjfkrIO3oS2bNXYW9ZfaKLQj0AW94UiZABrqqbAZz0m/xNAIuM24sAfCvK5UpaHxsjWDYf4thiIkqscPvAR6hqpXG7CsCIYDOKyEwRyRWRXJstsaHnbu1w6BYlC3ahUCQi3omprqNigm4IquoCVc1R1ZzMzMxIXy4sNY0tmL10D9oczoS8PhFRLGSE+bxqEclS1UoRyQKQ1EfG/H5lIVbtrcSAvuG+XaLYYlc4hSPcFvhKAHcat+8EsCI6xYmN7h6mXNfcFqOSEBFFj5lhhO8C2AbgyyJSJiJ3AZgL4BYROQTgZuN+j3Gwuslzm6MEiChZhexTUNU7gjx0U5TLEnvunZjccUREPUBKdApz1AnF25OrCjF17HDzT+CWHoUhJQKcKN5e3nIEL285kuhiUA+XUudCiWcjR1XxzLoDOGacAZGIKNpSI8CNHpR4XsjhyInT+PuGYvy/N3Lj9ppkXbygA4UjNQI8AdyrYzsPHiKiGGGAxwiHH1KiZc9ehZ+8lvyXk6PwMcBjjQNgyIRY/eBvOsCTrvVkKRHg7gxVv/uxxSY4EcVWSgR4IrEBTkSxklIBHs9+afaBE1GspUSAi3HsvJmhWtEeaig8bp9M4O+9de2vakD27FU4WN0Y99dOiQBPBK6QRKlh1R7XtW3WFlTF/bVTIsDdrWoz50SJVgO84+o/RESxkRIBTpTs4nmUMMVGIj7ClAjw7vSBExEFo6pwOgPnyCtbDse5NCkS4P662q8Y7YjnPszk9Jv3d3v6LhOJDW9rmb10Ly58eHXAxxpb7XEuTYoFeDgrS7gBzNZ+cvsgrwyz3tmV0DJUN7Rg+5HahJaBuue93OOdpiWyjZYS5wP3PxKzK1EfRgjBn1YX4aXNh1E6d0ZUl03WNmPeVpxoak10McjCUqoFbka04tv7d+ClzfHvG6Pkx/CmSDHA/US7T5J94EQUKxF1oYhIKYBGAA4AdlXNiUahok38+lC6Gg8erb5r7pyi7uDXhcIRjT7wG1T1RBSWExPPrj+IFfkVpudn8BKRWT94ZTtqm9oS9vo9vgvlpU9LEvK6gVry//yiHK12RwJKQ0Sx8O/iWuyviv85UNwiDXAFsE5E8kRkZjQKFEuJHtp333v5eHb9wYSWoadzOBWvbDmMlnZr/VDGcsuvoaU9dgu3mJZ2B2bM24LPS09GvKxkOHo20gC/RlUnA/g6gFkicp3/DCIyU0RyRSTXZkvs1UHM1HesPxNbA0cexNLyL8rxx1VFmPfJoUQXJWnUNLQkughJ41B1E/ZVNGDOyn0RLysJ8juyAFfVcuN/DYDlAKYEmGeBquaoak5mZmYkLxcX0W6l83Sy8dXc5joarrEl/kfFJS9+B93c63daFNbLJMjv8ANcRPqLyAD3bQC3AiiIVsESJdSvqtmPPRl+nVNZVz/EqbYfgm2IDu7TmKT1kDqJZBTKCADLjRZmBoB3VHVtVEoVI6aOxIzCMrx1+p70kC9OsjJTvXlHT8W8HJScnJ7zPEehBZ4ErbSwA1xVDwO4PIpliZ8EhqiZc5JT5Lpat1LhMyiuaUp0EZKSRrEFnvj4ToFhhN39oQ31qyom5wv39Sn2kuEziXXrbdmuspgu36rc9R6VPvAkSPAeH+DdreRQs7s/91DLDfZ4MoRHqkuGj2BrcWyPfUuCbElK7j7wZPgOREOPD3BvZlo9oXdiuj76k81tKCivD7k8/8DuKV+cZLIivxwVdWdcd0z8QibDyKA2uzPRRUhJnssrRvgV2FdRn/DjSoAUC3BTQnwm7r6z1/5div/4+9YuFpP4DzcVtNoduHdxPr63YLvP9K5qPwnyOyqb8F3ZVsLzjAfi/l5E+iM+Y95WdqHEQ7f7wEMEbzK03qiDw9gmtjW6DpAy8+kkxScY40LkH6+L7QtYlHsUigB4d+cxXPL42qCXSGtpdyTFSJOu9PgA967/aHwUkeY38z+67MbKl96NYQXJ8BnEswi90nr8am6eZxSK4Hcr9uF0mwPtzs7dWc1tdox/bC2eXncg6KLaHYnvBuMn68dsH7i/vKMn8eGezmc99J//dFtqHUQSa2qsQ90bFpa4BD9Q1YgN+6vjuiWXkZ4Ev1hJwnMgT4jkcx/J+35uGZ7fWIzs2as8W3v+y0qkHhXgTqfiyVWFKHfv0IJva0tN7IE2OwrF3+3zt+Ged77o9Fr+KrzKRpFzqG8L3H2QTpfjwBOYZ7f9bTN++npuXH9CzOZMed0ZXP/XjT36O9rRhSJddpe6vyNOBf7XOK+Of4s7GbpXelSAf3G8Di9vOYL7F+eHvQyz48DNSobN9Z7Mbmz+ugN8+RflIZ8T6x2IZiRDGfwt3nkMpbXN+CCv+2PI95TVobI++YO/YyemuYO9VNXr2A+/ZSU+v3tWgLsF6tMyK9RnkmZyWz3YcpLhQ+9JnJ4uFP/PpYvWVYBprXYHZr2zC0drT0etbF2JZ353t6UY+hgHRfbsVXhlS8e1Xr/xj3/j6rkbwileXDkDHMgT6P2mScf8nmM/4N+FkviVuUcFeJrXZo9ZK3dX4MZnNnm+5N39TJbmlSF79iqfaaqK+ZuKAQBFlQ2+j3Vv8RSCuwslI8AP65q9laaX81lJLVbtqcTjKyI/zagZce1CCfClW1tQhbJTzT7TzJbJvX49uboo5OsEM3XuBvzo1R3mn9CFXy/Jx5vbSk3N617PnaqeHeCBuPdRnGpuR7sjcDawDzzK3JWuqth0oAY1DS1oae/cGvfegfTAknwctp1Gm8k9yv4tvQfe391pnu2HT+KjfdUA4Pnwo0VV0cgT9Hu4h4BV1Hc+5/Uv3t7VaT67w9nlihs3ESa43eFE3tHAFyXwn34gwBVj7n4rDzPmBT6OIVjfcGFFA46c6NhCUQV+vHCn2SL7KK87gy2HonM06rJd5XjMxA9vQ0s7zrS51nPvc8UE+uF5aNkez233zssXjEaZ53lJ0ByzbIA7nNrpqivudcKpip+89jm+89K2gM894/W8NE/ou+6HHgfedbnsDmeXpyuNtOX17s7juHTOOp8Vyax2hzPomFezfvFWXsDRNtFy/GQzTreaO5f3YVsTfvtB5x9QoPNKeeHDq1FU2YDrn96E2+d/Fmkxg1LVgJ+//3c10j7wv647gNvnb+t0NHD9mXbcPt/3e+/fUvae14ffuuDtjx8WYvq8Lbjh6U0+0zcftMW077vN7sSK/PKIdxjur2rAZXPWYdY7rh917+tYBlrn3Q0wb89v9L08YxL0oFgnwPdV1GPl7o7g+O37uzH+Md+z17pXCrvR6j1a67uJ6PbYPws6PcfZkeA+/McXh1rtxj6ypssPNtLP/KN9VQCA0jACfNwja/CzN3Ijev01BVU+o22iqbzuDK79y8ZOR1UGc+/ifGw/HLgVuvjz452m5R+vQ9mpwGET7ND2Q9WNuHfxF7CH2EKzNbZi2t82Y+rcDfjyo53Pqvz3Db5XCDLTf7rpQI3PiCpVxeelJ6GqKKp0taptTR1XeMo7ehJrCzp3G/m/VrAf/2Df7dOtdryy9Yjn/op83x3FV/0peN93S7vD80NxsLoRb2wrDTqvv2O1zfjL2v24d3F+l+OxzZj2ty0+9723uCc8/hH+tbv7jRL/YYVdidWIFcsE+Ix5W/Grd13B4XQqlhmjDbwrxt2oMdsdAnT0m/+PEUr+1dw3I91v/tAtp0K/fm8fXuWtb27Hi5+WdPpwW9odeHJVIZpa7ahuaMH+qgZPf6VnJ4zXD0ur3YGmVjuKaxrR5Nd6Xbm7ApsO1Hjub9hf06nPPpCdRwKPa4+GpXll+Kyk8+bzVGMn2N4Q55gpKK9HXXNbyPn8PbRsb6dpb+84CgD485r9AIBPD9qwYHMJzrQ5MHvpHtzy3GasyK/wuXDtJ0XVmPTEOpw63dGKmzp3A/ZXNQbsygGAJr8rBH3/5dD9vz957XNMe24zjp9sxq+X5OOf+eX4zovb8H5uGTYfdF2esL65oxV9+/xteHBp5/fobsgcq23GC5uKfVrRga6X6b8OXPK7j3zu/3pJ4K0eoOOIWLfxj63F5b9fBwC49bnNeHzFPp8fw4eX+5b32XUHkHf0FKrqW3DdXzd6fjj8W7/uPnz3FZgi9T/vfoE7TDYc3AL9CB+rbUb27FW4/718/Gl1EZxOxZ/X7scFD60O2QgIRyQXdEiIjwurfQ4T/qykFlPHDgcAtBqtqMO20K3T2+d/hhd+MNlzYM0nRrBdPWaYz3y9MtJw6nQbluQexw+vHG1q9MBfPwreWthdVg9VxUf7qnD3W67Nucyz++Drl45En4x0VDW04IPcMry85Qhe3nLE57n/nn2jZyzqCWNFsTucPi2+CVkDsXzW1cg7egovbCzxnPVu3f2dLlfqY31hNbIG9UVLuwN/+LAQu8tc4bgyvwL/+P5k9M5I8+l+cTrV50fE6VQUVNQj/3gdHl+xD0VPTENGumDB5sO4dcIIDO3fG8PO7oO1BZWe/QYlT01Heprg1+/lY12h7yZrY0s7BvTt5bl/58Kd+PSgDSVPTQ96DppwWjmPLC9A/rE6HPZqlT61ej+eWr3fZ76mVjuyZ6/C2z/7Kh5ZXoBTze2Y9If1KJ07Awu3HgnZaFi07SgWbTva5Twrd1cgXQRXjxmGSX9YDwBobLXj0X8W4NODNizb5Wq0lHqNlLnvvXzc914+rhg9pMtlnzrdhuv+urHT9MvmrMOOh29Cia3J8+O/bFcZeqUJZt0w1rOT2KyvPPkxnvnO5bj9ivN8pj/xr0LP7b9v6OhLfmfHMVx0ztn4ydQLAADzNhRj3gbfvma3lnYH+vZyNajufiuv0+Otdgf6GA2uLYdsWFtQhYenX9y5qyiIbYdd548xu3M1UNW469g9nPU/LjsX8ze5fnxOtzowqF9028wSz8HoOTk5mpvb/U34gvL6Lk8cRUSU7G67ZARe+lFOWM8VkTxV7fRkS3ShMLyJyOpCbSWFwxIBTkRkdT+75sKoL9MSfeClc2cE3PE2IWsgHplxMc4f2g9rCiqxvrAan5d2XLD2tktG4NEZE/DcxwdRVNmIF34wGW9sK8WIgX0xd83+Tstzu2n8OZg8eohPX/bz35+M7OH90NzmwPMbi5E1qC/+e+oFONHYinWF1Xj9s1LPvHdMGYV3d3YeBfHv2Tdi55FanDOgL+ZvcvVPZw/rh4YWO1bMmoqn1x1A34x0zPnGJdh4oAa/NMYx33fzOPzt445RDF+7KBOHqn13mM35zwk40dSGf2z07T+8/+aL8NzHBzHjsixAgVV7K5H36M14c/tR9O+dga3FJ/DpQRumZA9Fm8OJ/ON1uGbscE/feZ+MNM++hfOH9sMDt16E+jPtWF9YjXtuGIuyU2dQ1dCC8rozeGfHsaB1CgDTLx2J9LQ0zx7/V36cg69eOBQvbCrx9BO++MPJnn0DAHD5qMF4664pEBG8n3scE0cNxrdfcA0DPKtXumdI6Lw7Jnl2cv/k6mwUVjbgvpvHoXd6Gs4dfBb698nw7Ey7OGsglv/yavz8zTx8etCG9+++CsP698bHRdU4UNWEO6aMQomtCQ8u3YsrRg/BfTePw8nTbbh3cT6+/9XzMe2SkViRX4GlxmXL+vVOxy++Nga3X3EePsgrw43jz8HYc87G4p3HMOdfhejbK63T8QjbHroRIwf2xbs7j+Ph5Xux6KdT0K93Or7zYuehr/fffBFKa0/jqguHYXdZHarqW3D9+HN8RlP5K507w3P7aO1p9MlIx9JdZZ7v9KTzB+PZ707E9sO1qKg7g1snjISI79buozMuRsOZdpScOI1+vdLx869diOKaJizcWorMgX3ww6+Oxleyh+CdncdQUdeCYf174+l1BzB6WD+cPN2GE15D9by9edcU7K9sxDs7j+HWCSPw0mbXEZ2/vuUifF560md8+ENfH4/3co/jK6OH4j8vPxd/+Wg/Rg/rj5svPgd/XFWEa8cOx7IvytG3VxpmXHoufnpNNjYU1cDW1IqdR07CqYqD1R1jvmfdMAZTxw7HsP59MGroWXhz21FcNWYYfvp6LmbdMAaTzh+CEQP7YM3eKnxWUouPizr2zbx111fx+melGD2sH17degSPzrgYg87qhRvGn4N73tnVaUSUe92bdslIXDZqkOmjuLvDEn3gAHDHgu3YdrgW6++/DpX1LSioqMcvrx8bdlncPwi7HrsFxTVNOHLCtcKOHNgX2x++CXaHE0tyy/DdnPOQkW5uQ6W2qRUZaWkY1K8XTrfaccUf16Ol3Yl/3XMNNh6owa9uGueZV1WxraQWV40ZFvTMdA0t7WizOzH87D6ob27H7S9+huKaJnz62+sxelh/2B1OKIBeXuXLLT2J/n0yMPzsPrA7ncgadBba7E70zujexlZjSzsunbMOMy7NwvM/mGz6eeV1Z7Dk8+P41U3jkJ4msDucSE+TkGff+9GrO7Dl0AksnnklrrxwGHYfr8Nl5w0K+bzimkY0ttgx6Xxzm6e7jp3CmMyzMeisXqFnDuHav2zA8ZNnfMIyGIdTkZ4mnu9doOfUNLRgylOfAAAKn7gNEx7/CONHDsDa+wLvgHbX2YiBfVDd0Bp0ud2RPXsVJmQNxOp7r41oOW4nmlpx5MRpfCV7aFSWF0hVfQtGDuob9PG65jYM7NsLbQ6nZyeoWUvzynDTxedgcL/epuZ3f759MtKwd85t3V7vggnWBw5VDfsPwDQABwAUA5gdav4rrrhCw3Wmza5lp5rDfr6/wop63XrI5rnvdDr148IqtTucUXuNosp6/ceGQ1FbnqqrnPFysKpBz7TZ4/JalXVn9KlVheqIYv3HWmNLu1bXn+nWc3JLa/W1rYcDPtbSbtfRD36ob2wrVVXVtQWVWtPQEnRZdodTD1Y1aN3pNh394Ic6+Yl13SpLIIUV9VrX3BbxclLV6Ac/1AsfWhX15QLI1QCZGnYLXETSARwEcAuAMgCfA7hDVQuDPSeSFjgRBZd/vA7nDu6LcwYEb4lS7L2y5TCuHZeJL48cENXlBmuBR9IHPgVAsaoeNl5gMYBvAgga4EQUGxNHDU50EQjAz66N/o7KrkTSQfMlAN576sqMaT5EZKaI5IpIrs1mi+DliIjIW8yHEarqAlXNUdWczMzMWL8cEVHKiCTAywGM8rp/njGNiIjiIJIA/xzAOBG5QER6A/gegJXRKRYREYUS9k5MVbWLyD0APgKQDmChqsbnciZERBTZkZiquhrA6iiVhYiIuoHnQiEisigGOBGRRcX1XCgiYgPQ9VntgxsOIDpXQe3ZWE/msJ5CYx2ZE496Gq2qncZhxzXAIyEiuYEOJSVfrCdzWE+hsY7MSWQ9sQuFiMiiGOBERBZlpQBfkOgCWATryRzWU2isI3MSVk+W6QMnIiJfVmqBExGRFwY4EZFFWSLARWSaiBwQkWIRmZ3o8sSbiJSKyF4RyReRXGPaUBFZLyKHjP9DjOkiIvOMutojIpO9lnOnMf8hEbkzUe8nWkRkoYjUiEiB17So1YuIXGHUe7Hx3OhflTYOgtTTHBEpN75T+SIy3euxh4z3fEBEbvOaHnA9NE5ot8OY/p5xcjtLEZFRIrJRRApFZJ+I3GtMT+7vU6DrrCXTH1wnyioBcCGA3gB2A5iQ6HLFuQ5KAQz3m/YXGNchBTAbwJ+N29MBrAEgAK4EsMOYPhTAYeP/EOP2kES/twjr5ToAkwEUxKJeAOw05hXjucRjSoQAAALVSURBVF9P9HuOYj3NAfCbAPNOMNaxPgAuMNa99K7WQwBLAHzPuP0igF8k+j2HUUdZACYbtwfAdbnICcn+fbJCC9xz6TZVbQPgvnRbqvsmgEXG7UUAvuU1/Q112Q5gsIhkAbgNwHpVPamqpwCsh+ui1JalqpsBnPSbHJV6MR4bqKrb1bX2veG1LEsJUk/BfBPAYlVtVdUjcF2wfAqCrIdGK/JGAB8Yz/euc8tQ1UpV3WXcbgRQBNcVxpL6+2SFADd16bYeTgGsE5E8EZlpTBuhqpXG7SoAI4zbweorVeoxWvXyJeO2//Se5B5j83+hu2sA3a+nYQDqVNXuN92yRCQbwCQAO5Dk3ycrBDgB16jqZABfBzBLRK7zftD4Red4UD+sly7NBzAGwEQAlQCeSWxxkoOInA1gKYD7VLXB+7Fk/D5ZIcBT/tJtqlpu/K8BsByuzdlqY7MMxv8aY/Zg9ZUq9Riteik3bvtP7xFUtVpVHarqBPAyXN8poPv1VAtX90GG33TLEZFecIX326q6zJic1N8nKwR4Sl+6TUT6i8gA920AtwIogKsO3Hu47wSwwri9EsCPjb3kVwKoNzYBPwJwq4gMMTaXbzWm9TRRqRfjsQYRudLo5/2x17Iszx1Khm/D9Z0CXPX0PRHpIyIXABgH1863gOuh0SrdCOC/jOd717llGJ/xqwCKVPVZr4eS+/uU6L2/JvcQT4drr3AJgEcSXZ44v/cL4drjvxvAPvf7h6vv8RMAhwB8DGCoMV0APG/U1V4AOV7L+ilcO6WKAfx3ot9bFOrmXbg2/9vh6lO8K5r1AiAHrmArAfAPGEcuW+0vSD29adTDHrjCKMtr/keM93wAXiMlgq2Hxnd0p1F/7wPok+j3HEYdXQNX98geAPnG3/Rk/z7xUHoiIouyQhcKEREFwAAnIrIoBjgRkUUxwImILIoBTkRkUQxwIiKLYoATEVnU/we+4eX1RLd1IAAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# Selecting which observations to remove based on the criteria\n","observations_to_remove = data_cleaned[data_cleaned['AveBedrms'] > 2.5]\n","data_cleaned = data_cleaned.drop(observations_to_remove.index)\n","print(f'Removing {len(observations_to_remove)} observations.')\n","observations_to_remove.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"-2ufK20wLvNR","executionInfo":{"status":"ok","timestamp":1639534114218,"user_tz":-60,"elapsed":734,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"0d21903a-da0c-46a4-d901-4cc1ac7f412d"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Removing 12 observations.\n"]},{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>710</th>\n","      <td>2.4196</td>\n","      <td>26.0</td>\n","      <td>8.518248</td>\n","      <td>2.700730</td>\n","      <td>253.0</td>\n","      <td>1.846715</td>\n","      <td>37.68</td>\n","      <td>-122.08</td>\n","      <td>2.750</td>\n","    </tr>\n","    <tr>\n","      <th>1925</th>\n","      <td>4.9583</td>\n","      <td>10.0</td>\n","      <td>5.062500</td>\n","      <td>2.562500</td>\n","      <td>55.0</td>\n","      <td>3.437500</td>\n","      <td>38.87</td>\n","      <td>-120.50</td>\n","      <td>0.875</td>\n","    </tr>\n","    <tr>\n","      <th>2776</th>\n","      <td>3.1607</td>\n","      <td>18.0</td>\n","      <td>9.314961</td>\n","      <td>2.858268</td>\n","      <td>374.0</td>\n","      <td>2.944882</td>\n","      <td>33.07</td>\n","      <td>-114.98</td>\n","      <td>0.575</td>\n","    </tr>\n","    <tr>\n","      <th>3334</th>\n","      <td>4.1250</td>\n","      <td>30.0</td>\n","      <td>5.384615</td>\n","      <td>2.923077</td>\n","      <td>20.0</td>\n","      <td>1.538462</td>\n","      <td>38.92</td>\n","      <td>-122.65</td>\n","      <td>1.125</td>\n","    </tr>\n","    <tr>\n","      <th>12307</th>\n","      <td>2.9141</td>\n","      <td>11.0</td>\n","      <td>9.855372</td>\n","      <td>2.731405</td>\n","      <td>682.0</td>\n","      <td>2.818182</td>\n","      <td>33.62</td>\n","      <td>-116.74</td>\n","      <td>2.143</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["       MedInc  HouseAge  AveRooms  ...  Latitude  Longitude  MedHouseVal\n","710    2.4196      26.0  8.518248  ...     37.68    -122.08        2.750\n","1925   4.9583      10.0  5.062500  ...     38.87    -120.50        0.875\n","2776   3.1607      18.0  9.314961  ...     33.07    -114.98        0.575\n","3334   4.1250      30.0  5.384615  ...     38.92    -122.65        1.125\n","12307  2.9141      11.0  9.855372  ...     33.62    -116.74        2.143\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":14}]},{"cell_type":"markdown","source":["###### Population feature"],"metadata":{"id":"-XB1o0MhN6nw"}},{"cell_type":"code","source":["data['Population'].plot()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"GPnnjb4CJjhI","executionInfo":{"status":"ok","timestamp":1639534114218,"user_tz":-60,"elapsed":10,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"b210772f-51db-4e9c-f1f5-8584042595c9"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7fb98507f050>"]},"metadata":{},"execution_count":15},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# Selecting which observations to remove based on the criteria\n","observations_to_remove = data_cleaned[data['Population'] > 15000]\n","data_cleaned = data_cleaned.drop(observations_to_remove.index)\n","print(f'Removing {len(observations_to_remove)} observations.')\n","observations_to_remove.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":224},"id":"5BJ_nT6ILvSR","executionInfo":{"status":"ok","timestamp":1639534114219,"user_tz":-60,"elapsed":9,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"f0839727-59b1-4bd9-9bf7-5ffc7467d563"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Removing 6 observations.\n"]},{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","      <th>MedHouseVal</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>6057</th>\n","      <td>6.0191</td>\n","      <td>8.0</td>\n","      <td>6.347327</td>\n","      <td>1.047525</td>\n","      <td>15507.0</td>\n","      <td>3.070693</td>\n","      <td>34.03</td>\n","      <td>-117.78</td>\n","      <td>2.539</td>\n","    </tr>\n","    <tr>\n","      <th>6066</th>\n","      <td>6.6288</td>\n","      <td>7.0</td>\n","      <td>6.802554</td>\n","      <td>1.026277</td>\n","      <td>15037.0</td>\n","      <td>3.692780</td>\n","      <td>34.04</td>\n","      <td>-117.87</td>\n","      <td>3.397</td>\n","    </tr>\n","    <tr>\n","      <th>9880</th>\n","      <td>2.3087</td>\n","      <td>11.0</td>\n","      <td>5.364518</td>\n","      <td>1.059684</td>\n","      <td>28566.0</td>\n","      <td>4.696810</td>\n","      <td>36.64</td>\n","      <td>-121.79</td>\n","      <td>1.188</td>\n","    </tr>\n","    <tr>\n","      <th>10309</th>\n","      <td>7.4947</td>\n","      <td>4.0</td>\n","      <td>7.311043</td>\n","      <td>1.054346</td>\n","      <td>16122.0</td>\n","      <td>3.106957</td>\n","      <td>33.89</td>\n","      <td>-117.74</td>\n","      <td>3.663</td>\n","    </tr>\n","    <tr>\n","      <th>13139</th>\n","      <td>4.9516</td>\n","      <td>3.0</td>\n","      <td>7.338559</td>\n","      <td>1.159015</td>\n","      <td>16305.0</td>\n","      <td>3.043113</td>\n","      <td>38.43</td>\n","      <td>-121.44</td>\n","      <td>1.537</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["       MedInc  HouseAge  AveRooms  ...  Latitude  Longitude  MedHouseVal\n","6057   6.0191       8.0  6.347327  ...     34.03    -117.78        2.539\n","6066   6.6288       7.0  6.802554  ...     34.04    -117.87        3.397\n","9880   2.3087      11.0  5.364518  ...     36.64    -121.79        1.188\n","10309  7.4947       4.0  7.311043  ...     33.89    -117.74        3.663\n","13139  4.9516       3.0  7.338559  ...     38.43    -121.44        1.537\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":16}]},{"cell_type":"markdown","source":["###### Overview of the cleaned dataset"],"metadata":{"id":"VDhBgoZ1OJTc"}},{"cell_type":"code","source":["# Original dataset\n","data.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uLM2CFUDLvXI","executionInfo":{"status":"ok","timestamp":1639534114219,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"1efbeb05-eab2-4cf6-a975-11e9f43a1136"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(20640, 9)"]},"metadata":{},"execution_count":17}]},{"cell_type":"code","source":["# Cleaned dataset\n","data_cleaned.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"95SgfNfzLDg8","executionInfo":{"status":"ok","timestamp":1639534114219,"user_tz":-60,"elapsed":6,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"8de425ae-4574-4557-9079-4a8479285317"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(20345, 9)"]},"metadata":{},"execution_count":18}]},{"cell_type":"code","source":["# Number of removed observations\n","print(f'{round(100 * (data.shape[0] - data_cleaned.shape[0]) / data.shape[0], 2)} % of data has been removed as outliers.')\n","print(f'{data.shape[0] - data_cleaned.shape[0]} observations have been removed as outliers.')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"CU75neS8Osb6","executionInfo":{"status":"ok","timestamp":1639534114219,"user_tz":-60,"elapsed":5,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"1ed845d7-f351-4ab6-dafc-5f4c465d0afc"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["1.43 % of data has been removed as outliers.\n","295 observations have been removed as outliers.\n"]}]},{"cell_type":"markdown","source":["Let's look now Box plots once again, to see how the feature value distribution has changed after removing values:"],"metadata":{"id":"M7r7gassSxuv"}},{"cell_type":"code","source":["plt.figure(figsize=(20, 40))\n","\n","for i, column in enumerate(data_cleaned.columns):\n","    # Create subplots\n","    plt.subplot(11, 2, i + 1)\n","    data_cleaned[column].plot.box(vert=False)\n","    # Labels and graphics\n","    plt.xlabel(None)\n","    plt.ylabel('Box plot of features')\n","    plt.title(column)\n","    plt.grid(color='white', linestyle='-.', linewidth=0.5)\n","    ax = plt.gca()\n","    ax.set_facecolor('lightgrey')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"id":"Ev5e5svwOslH","executionInfo":{"status":"ok","timestamp":1639534118030,"user_tz":-60,"elapsed":3815,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"76e5c024-6849-47cd-a6b0-657c4575a21a"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 1440x2880 with 9 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["data_cleaned.hist(figsize=(16, 12), bins=40, edgecolor='black')\n","plt.subplots_adjust(hspace=0.3, wspace=0.2)\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":716},"id":"l3yA-Gi6Qhgr","executionInfo":{"status":"ok","timestamp":1639534121990,"user_tz":-60,"elapsed":3962,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"5046bd14-11bf-4603-a37a-904733754434"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 1152x864 with 9 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["#### OLS regression on the original dataset"],"metadata":{"id":"PG59VtocTkoc"}},{"cell_type":"code","source":["# Evaluate the models using cross-validation\n","model_california_housing = make_pipeline(StandardScaler(), LinearRegression())\n","cv_california_housing    = cross_validate(model_california_housing, \n","                                          data.drop(columns=['MedHouseVal']), data['MedHouseVal'],\n","                                          return_estimator=True, n_jobs=1, cv=10)"],"metadata":{"id":"Mf5Drt5_Twd_"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Get R-squared\n","score_california_housing = cv_california_housing['test_score']\n","print(f'R2 score: {score_california_housing.mean():.3f} +/- {score_california_housing.std():.3f}')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UqsXV2kJTwjh","executionInfo":{"status":"ok","timestamp":1639534122537,"user_tz":-60,"elapsed":5,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"99ed0038-bbb2-49cc-ada8-5acc74c13a02"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["R2 score: 0.511 +/- 0.059\n"]}]},{"cell_type":"markdown","source":["#### OLS regression on the cleaned dataset"],"metadata":{"id":"OHTEwxkATse4"}},{"cell_type":"code","source":["# Evaluate the models using cross-validation\n","model_california_housing = make_pipeline(StandardScaler(), LinearRegression())\n","cv_california_housing    = cross_validate(model_california_housing, \n","                                          data_cleaned.drop(columns=['MedHouseVal']), data_cleaned['MedHouseVal'],\n","                                          return_estimator=True, n_jobs=1, cv=10)"],"metadata":{"id":"duu89sTtTvtU"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Get R-squared\n","score_california_housing = cv_california_housing['test_score']\n","print(f'R2 score: {score_california_housing.mean():.3f} +/- {score_california_housing.std():.3f}')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Tx-KIBF0Tvxu","executionInfo":{"status":"ok","timestamp":1639534123026,"user_tz":-60,"elapsed":18,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"9cd3e27a-1758-4209-f8ec-c5a863b576bc"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["R2 score: 0.580 +/- 0.048\n"]}]},{"cell_type":"markdown","source":["#### Model comparison"],"metadata":{"id":"Ldv3VCEPWqQB"}},{"cell_type":"markdown","source":["We got that the model fitted on the cleaned dataset using OLS regression has a higher R-squared score.\n","\n","|            | OLS (original dataset) | OLS (cleaned dataset) |\n","|------------|------------------------|-----------------------|\n","| R-squared  | 0.511 +/- 0.059        | 0.580 +/- 0.048.      |"],"metadata":{"id":"qHKfUiQqWsz3"}},{"cell_type":"markdown","source":["### (b) Feature and model selection:"],"metadata":{"id":"5BnfSfS68oI5"}},{"cell_type":"markdown","source":["#### (b.1) Using your final (transformed) dataset, split the full sample into training and testing subsamples. Evaluate in-sample and out-of-sample performance for the following models: OLS, PLS, rigde, lasso, and elastic net regressions. For models which do not embed a variable selection procedure, consider specifications with different numbers of features. You can use performance metric(s) of your choice, however provide a rationale for your preferred approach.\n"],"metadata":{"id":"eTztaurA8hKj"}},{"cell_type":"markdown","source":["In order to protect from overfitting our model, we're goind to perform feature selection or adding regularization terms to our models.\n","\n","Our dataset contains just over 20,000 observations and 9 features. The number of observations is much larger than the number of features, so we don't expect dimensionality problems nor multicolinearity problem amoung features in dataset as we previously seen that the corrlerations are relatively low."],"metadata":{"id":"gdegAXG27pY8"}},{"cell_type":"code","source":["data_cleaned.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"gxCLBXir79Qy","executionInfo":{"status":"ok","timestamp":1639534123026,"user_tz":-60,"elapsed":12,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"943176ab-932b-4fbd-f6d6-4b3ca8fe35d3"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(20345, 9)"]},"metadata":{},"execution_count":26}]},{"cell_type":"code","source":["data_cleaned.columns"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"rGmGMfrnUVzm","executionInfo":{"status":"ok","timestamp":1639534123026,"user_tz":-60,"elapsed":9,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"7d198ca5-c0a4-4748-f616-962a62b5cbd8"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Index(['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population', 'AveOccup',\n","       'Latitude', 'Longitude', 'MedHouseVal'],\n","      dtype='object')"]},"metadata":{},"execution_count":27}]},{"cell_type":"markdown","source":["Before we move on, we'll standardize the features and target variable:"],"metadata":{"id":"VZI7jWGKOr1S"}},{"cell_type":"code","source":["# Scale features\n","scaler = StandardScaler()\n","\n","scaled_data = scaler.fit_transform(data_cleaned)\n","X = scaled_data[:,0:-1]\n","y = scaled_data[:,-1].reshape(1,-1)\n","\n","X.shape, y.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"CXH5ujxIO43G","executionInfo":{"status":"ok","timestamp":1639534123026,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"136c9fc1-ca7f-415d-894d-8114e2b896c4"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((20345, 8), (1, 20345))"]},"metadata":{},"execution_count":28}]},{"cell_type":"markdown","source":["Let's now also split our dataset into the:\n","- Training set (80% of the observations ~ 16,276 observations)\n","- Testing set (20% of the observations ~ 4,069 observations)"],"metadata":{"id":"5QsDLuG0QcmA"}},{"cell_type":"code","source":["from sklearn.model_selection import train_test_split\n","\n","X_train, X_test, y_train, y_test = train_test_split(X, y[0,:], test_size=0.2, random_state=42)\n","\n","print(f'Features: train={len(X_train)} test={len(X_test)}')\n","print(f'Target: train={len(y_train)} test={len(y_test)}')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"qQE5dEhaPo66","executionInfo":{"status":"ok","timestamp":1639534123027,"user_tz":-60,"elapsed":6,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"ae8ac08d-e9ca-4cb9-cc82-053b49ee8b7c"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Features: train=16276 test=4069\n","Target: train=16276 test=4069\n"]}]},{"cell_type":"markdown","source":["#### 1.OLS Regression"],"metadata":{"id":"Y4OSTUzlbDFB"}},{"cell_type":"markdown","source":["Vanilla OLS regression doesn't have embeded feature selection or any overfitting prevention methods, we'll manually select features to include in the regression model."],"metadata":{"id":"8G9i4UWZR1JU"}},{"cell_type":"markdown","source":["For feature selection, we'll consider univariate feature selection. Regressions will be performed between every single feature and target variable, and based on that we'll chose a subset of features to include."],"metadata":{"id":"DsYzkOFiTApq"}},{"cell_type":"markdown","source":["From the results down bellow we can see that based on two different metrics to evaluate univariate feature regressions, we can see that sets for K most important features is defferent. It's different for all different number of most important features (3-7 of them). Only when we select 7 best features we got the same set of features when evaluated with different metrics.\n"],"metadata":{"id":"g-nslcx-ajFF"}},{"cell_type":"code","source":["# SelectKBest\n","for n_select in [3, 4, 5, 6, 7]:\n","  # Performance metric: f_regression\n","  kbest_freg_california_housing  = SelectKBest(score_func=f_regression, k=n_select)  \n","  X_scaled_kbest_freg            = kbest_freg_california_housing.fit_transform(X, y[0,:])\n","  mask_kbest_freg                = kbest_freg_california_housing.get_support()\n","  feature_names_kbest_freg       = data_cleaned.columns[:-1][mask_kbest_freg]\n","\n","  # Performance metric: mutual_info_regression\n","  kbest_mireg_california_housing = SelectKBest(score_func=mutual_info_regression, k=n_select)  \n","  X_scaled_kbest_mireg           = kbest_mireg_california_housing.fit_transform(X, y[0,:])\n","  mask_kbest_mireg               = kbest_mireg_california_housing.get_support()\n","  feature_names_kbest_mireg      = data.columns[:-1][mask_kbest_mireg]\n","\n","  print(f'{n_select} features')\n","  print(f'SelectKBest (f_regression): {\", \".join(feature_names_kbest_freg)}.')\n","  print(f'SelectKBest (mutual_info_regression): {\", \".join(feature_names_kbest_mireg)}.')\n","  print('-'*20)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"koQ055jkT7v3","executionInfo":{"status":"ok","timestamp":1639534133825,"user_tz":-60,"elapsed":10802,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"fd4a1b25-ff7e-4f08-aa19-184536a2f9b0"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["3 features\n","SelectKBest (f_regression): MedInc, AveRooms, AveOccup.\n","SelectKBest (mutual_info_regression): MedInc, Latitude, Longitude.\n","--------------------\n","4 features\n","SelectKBest (f_regression): MedInc, AveRooms, AveOccup, Latitude.\n","SelectKBest (mutual_info_regression): MedInc, AveRooms, Latitude, Longitude.\n","--------------------\n","5 features\n","SelectKBest (f_regression): MedInc, HouseAge, AveRooms, AveOccup, Latitude.\n","SelectKBest (mutual_info_regression): MedInc, AveRooms, AveOccup, Latitude, Longitude.\n","--------------------\n","6 features\n","SelectKBest (f_regression): MedInc, HouseAge, AveRooms, AveBedrms, AveOccup, Latitude.\n","SelectKBest (mutual_info_regression): MedInc, HouseAge, AveRooms, AveOccup, Latitude, Longitude.\n","--------------------\n","7 features\n","SelectKBest (f_regression): MedInc, HouseAge, AveRooms, AveBedrms, AveOccup, Latitude, Longitude.\n","SelectKBest (mutual_info_regression): MedInc, HouseAge, AveRooms, AveBedrms, AveOccup, Latitude, Longitude.\n","--------------------\n"]}]},{"cell_type":"markdown","source":["Because we couldn't manually determine the number and which features to use for our model, based on univariate testing, now we'll try to determine the optimal number of features and which one to use. For that, Recursive feature elimination with cross-validation to select the number of features is used (from scikit-learn package).\n","\n","The minumum number of features we'll set to be 4, out of 8 features in total.\n","\n"],"metadata":{"id":"beC1olGubnGs"}},{"cell_type":"code","source":["# Backward stepwise selection using scikit-learn RFECV\n","n_min_select = 4\n","n_cv_rfe     = 10\n","\n","ols_estimator   = LinearRegression()\n","rfecv_estimator = RFECV(estimator=ols_estimator, min_features_to_select=n_min_select, step=1, cv=n_cv_rfe)\n","\n","rfecv_california_housing = rfecv_estimator.fit_transform(X, y[0,:])\n","mask_rfecv_california_housing = rfecv_estimator.get_support()\n","feature_names_rfecv_california_housing = data_cleaned.columns[:-1][mask_rfecv_california_housing]\n","\n","print(f'The optimal number of features using RFECV: {rfecv_estimator.n_features_}.')\n","print(f'The selected features using RFECV: {feature_names_rfecv_california_housing}.')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"l46606XbW90s","executionInfo":{"status":"ok","timestamp":1639534134311,"user_tz":-60,"elapsed":507,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"95b8ed2f-f013-468b-9f11-b8c0f16613e9"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["The optimal number of features using RFECV: 8.\n","The selected features using RFECV: Index(['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population', 'AveOccup',\n","       'Latitude', 'Longitude'],\n","      dtype='object').\n"]}]},{"cell_type":"markdown","source":["We got that we should use 8 features (all of them):\n","- Population\n","- AveOccup\n","- Latitude\n","- Longitude\n","- MedInc\n","- HouseAge\n","- AveRooms\n","- Avebedrms\n"],"metadata":{"id":"Zwtg70Xtd29J"}},{"cell_type":"markdown","source":["In the previous step we got that the optimal features to use is all. Now let's try to select the optimal 4 (half of them) features to use. We'll create and evaluate the model using both the best 4 features and all features.\n","\n","We're running Recursive feature elimination (RFE) to select features by recursively considering smaller and smaller sets of features. \n","\n","In the results bellow we got the the optimal 4 features to use are:\n","\n","- 'MedInc'\n","- 'AveOccup'\n","- 'Latitude'\n","- 'Longitude'\n","\n","All of that makes sense. Two columns are about location, and in real estate market the most important thing is location. AveOccup says how big the house is, how many poople live in it. And MedInc explains if it's a rich or poor neighbourhood"],"metadata":{"id":"ZOoUoB53oKfj"}},{"cell_type":"code","source":["# Backward stepwise selection using scikit-learn RFE\n","n_select = 4\n","\n","ols_estimator   = LinearRegression()\n","rfe_estimator = RFE(estimator=ols_estimator, n_features_to_select=n_select, step=1)\n","\n","rfe_california_housing = rfe_estimator.fit_transform(X, y[0,:])\n","mask_rfe_california_housing = rfe_estimator.get_support()\n","feature_names_rfe_california_housing = data_cleaned.columns[:-1][mask_rfe_california_housing]\n","\n","print(f'The selected {n_select} features using RFE: {feature_names_rfe_california_housing}.')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"iLOsVJr2nx-c","executionInfo":{"status":"ok","timestamp":1639534134312,"user_tz":-60,"elapsed":15,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"5325ca90-6351-431f-923f-b4a273c572ac"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["The selected 4 features using RFE: Index(['MedInc', 'AveOccup', 'Latitude', 'Longitude'], dtype='object').\n"]}]},{"cell_type":"code","source":["# Indexes of selected 4 features\n","feature_indexes = []\n","for i, col in enumerate(data_cleaned.columns):\n","  if col in feature_names_rfe_california_housing:\n","    feature_indexes.append(i)\n","feature_indexes, data_cleaned.columns"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-KvO6OsDVBhO","executionInfo":{"status":"ok","timestamp":1639534134312,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"ad65dc9d-161a-4375-f881-3a27fca9659c"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["([0, 5, 6, 7],\n"," Index(['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population', 'AveOccup',\n","        'Latitude', 'Longitude', 'MedHouseVal'],\n","       dtype='object'))"]},"metadata":{},"execution_count":33}]},{"cell_type":"code","source":["# Selected features for OLS regression\n","X_train_select = X_train[:, feature_indexes]\n","X_test_select = X_test[:, feature_indexes]\n","X_train_select.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"atHabR7AgLrI","executionInfo":{"status":"ok","timestamp":1639534134313,"user_tz":-60,"elapsed":10,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"5d20c5dd-e9fc-4603-84cc-22485325b371"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(16276, 4)"]},"metadata":{},"execution_count":34}]},{"cell_type":"markdown","source":["**In-sample evaluation**\n","\n","Let's now perform in-sample model evalidation using K-fold cross validation to find the optimal model parameters. We'll do regression model wiith all features and with a selected number of them.\n","\n","For the in-sample evaluation, we'll look at R-squared score and RMSE (or negative of root mean square error; using negation for consistency reason if finding the model with the highest score)."],"metadata":{"id":"JDbGVo0WhpMd"}},{"cell_type":"markdown","source":["1. Selected features:"],"metadata":{"id":"zSttqpmZrqO0"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","ols_model = LinearRegression()\n","scores_select = cross_validate(ols_model, X_train_select, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"v_ujHxw2g7H-"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["scores_select['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"9H1yiN9uicLq","executionInfo":{"status":"ok","timestamp":1639534134313,"user_tz":-60,"elapsed":8,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"6e5769c4-e275-41c5-9265-43aa9d445df2"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.56584846, -0.63115635, -0.56827794, -0.59079927, -0.59273371,\n","       -0.58972412, -0.57636591, -0.59229836, -0.61035616, -0.61984955])"]},"metadata":{},"execution_count":36}]},{"cell_type":"code","source":["scores_select['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"fS0pcQKyrisw","executionInfo":{"status":"ok","timestamp":1639534134313,"user_tz":-60,"elapsed":5,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"cb5dc4d7-50c0-49ca-8c86-6479170577ab"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.65206427, 0.60846636, 0.65669822, 0.6464273 , 0.64894946,\n","       0.66361487, 0.66464173, 0.65099599, 0.62956916, 0.6303831 ])"]},"metadata":{},"execution_count":37}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores_select['test_r2'], 'g-')\n","ax2.plot(scores_select['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('Regression models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"id":"2TMnP8mJsa9T","executionInfo":{"status":"ok","timestamp":1639534135093,"user_tz":-60,"elapsed":784,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"0c6bc760-bc8c-4161-e269-8fec597d1bc6"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["From the 10 different models built and evaluated in the cross validation, we can see that there isn't a signle model which is best for both metrics.\n","\n","We'll prioritize R-squared metrics over RMSE.\n","\n","So the best in-sample performing model is model 6!"],"metadata":{"id":"FzPcoGSDwcNz"}},{"cell_type":"code","source":["# Intercept and coefficients\n","scores_select['estimator'][6].intercept_, scores_select['estimator'][6].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"U6EttmTx1nxC","executionInfo":{"status":"ok","timestamp":1639534135094,"user_tz":-60,"elapsed":20,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"62c02eb7-f819-4da3-c8b6-80675b4be42a"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(-0.001215613579151392,\n"," array([ 0.58148   , -0.23291029, -0.93985135, -0.86906289]))"]},"metadata":{},"execution_count":39}]},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_ols_selected = scores_select['estimator'][6]"],"metadata":{"id":"LlI6Gx8X1sKW"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison = pd.DataFrame([], columns=['Model', 'R-squared (in-sample)', 'RMSE (in-sample)', 'RMSE (out-of-the-sample)'])\n","model_comparison.loc[0] = ['OLS (4 features)', scores_select['test_r2'][6], -scores_select['test_neg_root_mean_squared_error'][6], \"\"]"],"metadata":{"id":"qbNKAOBo26z1"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["2. All features:"],"metadata":{"id":"h0j8m_0Us2r5"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","scores_all = cross_validate(ols_model, X_train, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"fb4Hq-Crs87N"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["scores_all['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"_mWWjg21tEpz","executionInfo":{"status":"ok","timestamp":1639534135095,"user_tz":-60,"elapsed":12,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"c654522e-dba4-4825-eec5-157f4b60bf2d"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.67496562, 0.62342865, 0.66965402, 0.66914844, 0.67229613,\n","       0.68318712, 0.67891389, 0.67617007, 0.64784193, 0.65412038])"]},"metadata":{},"execution_count":43}]},{"cell_type":"code","source":["scores_all['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"8VXhkCrEtFdd","executionInfo":{"status":"ok","timestamp":1639534135095,"user_tz":-60,"elapsed":9,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"9d32f0b5-2ecd-42de-a66c-6cff6582c070"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.54690925, -0.61897919, -0.55745174, -0.57150125, -0.5726847 ,\n","       -0.57231076, -0.56396809, -0.57053699, -0.59511183, -0.59961552])"]},"metadata":{},"execution_count":44}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores_all['test_r2'], 'g-')\n","ax2.plot(scores_all['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('Regression models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"id":"BtCLenwptAtz","executionInfo":{"status":"ok","timestamp":1639534135739,"user_tz":-60,"elapsed":650,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"381b2c5a-f38b-4d05-95e4-e29ca5e42dc4"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["Here the model with the highest R-squared score is model 5. The Difference in RMSE between top performing models isn't so wide spread."],"metadata":{"id":"WL-RfsX04dWy"}},{"cell_type":"code","source":["# Intercept and coefficients\n","scores_all['estimator'][5].intercept_, scores_all['estimator'][5].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"eeYPcAzP5Gi9","executionInfo":{"status":"ok","timestamp":1639534135739,"user_tz":-60,"elapsed":16,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"c52f9f20-5b09-4779-e871-c72040ece2cb"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(0.0013703972047001724,\n"," array([ 0.70612631,  0.12250136, -0.10915646,  0.10647406,  0.04257997,\n","        -0.22210976, -0.80018055, -0.73693909]))"]},"metadata":{},"execution_count":46}]},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_ols_all = scores_all['estimator'][5]"],"metadata":{"id":"NYIvU-jd5OVj"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[1] = ['OLS (all features)', scores_all['test_r2'][5], -scores_all['test_neg_root_mean_squared_error'][5], \"\"]"],"metadata":{"id":"Q4rfvCM65Whp"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Out-of-the-sample evaluation**\n","\n","Now let's perform out-of-the-sample evaluation on our test dataset. For measuring performance of test set, we'll use RMSE."],"metadata":{"id":"Jkh7Z0AO5_VW"}},{"cell_type":"code","source":["def rmse(predictions, targets):\n","    return np.sqrt(((predictions - targets) ** 2).mean())"],"metadata":{"id":"ZMHqWEBQ7yjH"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["1. Model with 4 features"],"metadata":{"id":"QiAwX2hG62X0"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_ols_selected.predict(X_test_select)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"6veJxgrR67mS","executionInfo":{"status":"ok","timestamp":1639534135741,"user_tz":-60,"elapsed":14,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"082d0f2f-b581-4f22-9215-19c3c946894c"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.6080983700570384"]},"metadata":{},"execution_count":50}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[0, 'RMSE (out-of-the-sample)'] = error"],"metadata":{"id":"CZw8UzIE73v1"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["2. Model with all features"],"metadata":{"id":"Fuhxfbrq65cb"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_ols_all.predict(X_test)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uPMsyQNL8Ml3","executionInfo":{"status":"ok","timestamp":1639534135741,"user_tz":-60,"elapsed":12,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"c7ccfec3-e080-4487-f0a5-f88c9511ad6c"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.5860471752200825"]},"metadata":{},"execution_count":52}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[1, 'RMSE (out-of-the-sample)'] = error"],"metadata":{"id":"ws-rKGhe8Vxo"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Model comparison**\n","\n","As we can see from the table bellow, OLS model using only 4 features had a lower R-squared score in-sample, and out-of-the-sample it performed slightly worse (higher RMSE).\n","\n","All in all, OLS regression on all features was the best performing model."],"metadata":{"id":"R-faPvFy8YEg"}},{"cell_type":"code","source":["model_comparison"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":112},"id":"46OmEiNJ8aPs","executionInfo":{"status":"ok","timestamp":1639534135741,"user_tz":-60,"elapsed":10,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"96ddb299-a3e2-4b0a-a785-74cae38bd0ea"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>R-squared (in-sample)</th>\n","      <th>RMSE (in-sample)</th>\n","      <th>RMSE (out-of-the-sample)</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>OLS (4 features)</td>\n","      <td>0.664642</td>\n","      <td>0.576366</td>\n","      <td>0.608098</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>OLS (all features)</td>\n","      <td>0.683187</td>\n","      <td>0.572311</td>\n","      <td>0.586047</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                Model  ...  RMSE (out-of-the-sample)\n","0    OLS (4 features)  ...                  0.608098\n","1  OLS (all features)  ...                  0.586047\n","\n","[2 rows x 4 columns]"]},"metadata":{},"execution_count":54}]},{"cell_type":"markdown","source":["#### 2.PLS Regression"],"metadata":{"id":"58Zv2KOgbFbJ"}},{"cell_type":"markdown","source":["Partial least squares (PLS) regression is a technique that reduces the predictors to a smaller set of uncorrelated components and performs least squares regression on these components, instead of on the original data. PLS regression is especially useful when your predictors are highly collinear, or when you have more predictors than observations and ordinary least-squares regression either produces coefficients with high standard errors or fails completely.\n","\n","Because of embeded properties of PLS regression, that are similar to PCA, we won't be doing feature selection for this regression model."],"metadata":{"id":"JDYQL2f7A5Lk"}},{"cell_type":"markdown","source":["**In-sample evaluation**"],"metadata":{"id":"LwOdNITjCXQv"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","pls_model = PLSRegression(n_components = 4)\n","scores = cross_validate(pls_model, X_train, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"atTihzEJCn4N"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Cross-validation: R-squared\n","scores['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"nw2py3TmDrOB","executionInfo":{"status":"ok","timestamp":1639534135742,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"d707d788-5fc8-4c22-c45d-268159da17d5"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.65261083, 0.59638208, 0.63969426, 0.64772284, 0.6445598 ,\n","       0.65037066, 0.65119276, 0.65934829, 0.61692328, 0.62913278])"]},"metadata":{},"execution_count":56}]},{"cell_type":"code","source":["# Cross-validation: RMSE (negative)\n","scores['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"CNjkOL_8DwCA","executionInfo":{"status":"ok","timestamp":1639534135742,"user_tz":-60,"elapsed":8,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"becd08a0-bc8f-41d5-d280-3335f294bce0"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.56540384, -0.64082232, -0.58218144, -0.5897159 , -0.59642807,\n","       -0.60122141, -0.58780939, -0.58516805, -0.62068699, -0.62089706])"]},"metadata":{},"execution_count":57}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores['test_r2'], 'g-')\n","ax2.plot(scores['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('Regression models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"id":"SvQqDd5eEAQX","executionInfo":{"status":"ok","timestamp":1639534136637,"user_tz":-60,"elapsed":902,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"f4b3a1e6-4b5f-4fe9-e2c4-109a65d3ef8f"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["Following already established rules for in-sample model evaluation and selection, the best performing model is model 0. It has the second lowest R-squared score and the lowest RMSE."],"metadata":{"id":"uE0dRiBlEHR-"}},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_pls = scores['estimator'][0]"],"metadata":{"id":"Kb_cGsFgEUmu"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Intercept and coefficients\n","scores['estimator'][0].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"fgfSGcR5F59k","executionInfo":{"status":"ok","timestamp":1639534136638,"user_tz":-60,"elapsed":19,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"71b7bc4d-13d6-4764-dc92-a9ad52c1ab78"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[ 0.80882126],\n","       [ 0.1846303 ],\n","       [-0.17242599],\n","       [ 0.14307735],\n","       [ 0.12823572],\n","       [-0.23576525],\n","       [-0.39613419],\n","       [-0.30751056]])"]},"metadata":{},"execution_count":60}]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[2] = ['PLS', scores['test_r2'][0], -scores['test_neg_root_mean_squared_error'][0], \"\"]\n","model_comparison"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":143},"id":"in-dZ0TWFqr4","executionInfo":{"status":"ok","timestamp":1639534136638,"user_tz":-60,"elapsed":15,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"dbbaea10-ce70-41c6-dbe8-d7715e87e622"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>R-squared (in-sample)</th>\n","      <th>RMSE (in-sample)</th>\n","      <th>RMSE (out-of-the-sample)</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>OLS (4 features)</td>\n","      <td>0.664642</td>\n","      <td>0.576366</td>\n","      <td>0.608098</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>OLS (all features)</td>\n","      <td>0.683187</td>\n","      <td>0.572311</td>\n","      <td>0.586047</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>PLS</td>\n","      <td>0.652611</td>\n","      <td>0.565404</td>\n","      <td></td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                Model  ...  RMSE (out-of-the-sample)\n","0    OLS (4 features)  ...                  0.608098\n","1  OLS (all features)  ...                  0.586047\n","2                 PLS  ...                          \n","\n","[3 rows x 4 columns]"]},"metadata":{},"execution_count":61}]},{"cell_type":"markdown","source":["**Out-of-the-sample evaluation**"],"metadata":{"id":"KSb0s9NiGQXy"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_pls.predict(X_test)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5R13f9BZGWGR","executionInfo":{"status":"ok","timestamp":1639534136638,"user_tz":-60,"elapsed":14,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"3b0d798c-7ed9-45b3-dc4d-458c1d9bfb43"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["1.2723136557008117"]},"metadata":{},"execution_count":62}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[2, 'RMSE (out-of-the-sample)'] = error\n","model_comparison"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":143},"id":"_vnGJw7OJwIx","executionInfo":{"status":"ok","timestamp":1639534136639,"user_tz":-60,"elapsed":12,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"0d289b97-f47c-4803-f469-396ffefc6a61"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>R-squared (in-sample)</th>\n","      <th>RMSE (in-sample)</th>\n","      <th>RMSE (out-of-the-sample)</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>OLS (4 features)</td>\n","      <td>0.664642</td>\n","      <td>0.576366</td>\n","      <td>0.608098</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>OLS (all features)</td>\n","      <td>0.683187</td>\n","      <td>0.572311</td>\n","      <td>0.586047</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>PLS</td>\n","      <td>0.652611</td>\n","      <td>0.565404</td>\n","      <td>1.27231</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                Model  ...  RMSE (out-of-the-sample)\n","0    OLS (4 features)  ...                  0.608098\n","1  OLS (all features)  ...                  0.586047\n","2                 PLS  ...                   1.27231\n","\n","[3 rows x 4 columns]"]},"metadata":{},"execution_count":63}]},{"cell_type":"markdown","source":["Now we'll move to `regularized regressions`. They prevent overfitting by adding a panalization term to the loss function, which penalizes more complex models, similar like AIC and BIC score do."],"metadata":{"id":"lHAo4iNONPEC"}},{"cell_type":"markdown","source":["#### 3.Ridge Regression"],"metadata":{"id":"R8HCJx5ybH6A"}},{"cell_type":"markdown","source":["For Rigde regression increasing the hyperparameter  𝜆  shrinks the coefficients toward zero. For this reason,  𝜆  is often called `shrinkage intensity`.\n","\n","Unlike LASSO regression, [RIDGE regression does not have zeroing coefficients as a goal](https://stats.stackexchange.com/questions/151954/sparsity-in-lasso-and-advantage-over-ridge-statistical-learning/151960#151960), which means that `ridge regression does not help with the variables selection`. It is possible that some coefficients will be equal to zero, but that's\n","\n","Similar like for OLS regression, we'll first perform the feature selection process."],"metadata":{"id":"6vhkLccGOFCx"}},{"cell_type":"markdown","source":["At the beginning we tried to perform feature selection based on univariate testing, now we'll try to determine the optimal number of features and which one to use. As a consequecnce, recursive feature elimination with cross-validation will be used to select the number of features is used (from scikit-learn package).\n","\n","\n","The minumum number of features we'll set to be 4, out of 8 features in total.\n","\n","Similar as for the OLS, we got that we should use all features.\n"],"metadata":{"id":"FfARG2vDfI8W"}},{"cell_type":"code","source":["# Backward stepwise selection using scikit-learn RFECV\n","lambda_param = 100\n","n_min_select = 4\n","n_cv_rfe     = 10\n","\n","model_ridge = Ridge(alpha=lambda_param)\n","rfecv_estimator = RFECV(estimator=model_ridge, min_features_to_select=n_min_select, step=1, cv=n_cv_rfe)\n","\n","rfecv_california_housing = rfecv_estimator.fit_transform(X, y[0,:])\n","mask_rfecv_california_housing = rfecv_estimator.get_support()\n","feature_names_rfecv_california_housing = data_cleaned.columns[:-1][mask_rfecv_california_housing]\n","\n","print(f'The optimal number of features using RFECV: {rfecv_estimator.n_features_}.')\n","print(f'The selected features using RFECV: {feature_names_rfecv_california_housing}.')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"pgXfeMkce_hY","executionInfo":{"status":"ok","timestamp":1639534137264,"user_tz":-60,"elapsed":636,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"dfb1458b-2b42-408c-9d5e-64dc52552da8"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["The optimal number of features using RFECV: 8.\n","The selected features using RFECV: Index(['MedInc', 'HouseAge', 'AveRooms', 'AveBedrms', 'Population', 'AveOccup',\n","       'Latitude', 'Longitude'],\n","      dtype='object').\n"]}]},{"cell_type":"markdown","source":["In the previous step we got that the optimal features to use is all. Now let's try to select the optimal 4 (half of them) features to use. We'll create and evaluate the model using both the best 4 features and all features.\n","\n","We're running Recursive feature elimination (RFE) to select features by recursively considering smaller and smaller sets of features. \n","\n","In the results bellow we got the the optimal 4 features to use are:\n","\n","- 'MedInc'\n","- 'AveOccup'\n","- 'Latitude'\n","- 'Longitude'\n","\n","Those are the same features that we got for OLS regressing using backward stepwise selection."],"metadata":{"id":"ThaXUcA6gKd_"}},{"cell_type":"code","source":["# Backward stepwise selection using scikit-learn RFE\n","n_select = 4\n","lambda_param = 100\n","\n","model_ridge = Ridge(alpha=lambda_param)\n","rfe_estimator = RFE(estimator=model_ridge, n_features_to_select=n_select, step=1)\n","\n","rfe_california_housing = rfe_estimator.fit_transform(X, y[0,:])\n","mask_rfe_california_housing = rfe_estimator.get_support()\n","feature_names_rfe_california_housing = data_cleaned.columns[:-1][mask_rfe_california_housing]\n","\n","print(f'The selected {n_select} features using RFE: {feature_names_rfe_california_housing}.')"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ziJZwRRde_wD","executionInfo":{"status":"ok","timestamp":1639534137265,"user_tz":-60,"elapsed":14,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"b4f38cb5-93ad-45ce-dad1-91d0d1234057"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["The selected 4 features using RFE: Index(['MedInc', 'AveOccup', 'Latitude', 'Longitude'], dtype='object').\n"]}]},{"cell_type":"markdown","source":["**In sample evaluation**"],"metadata":{"id":"GL5sSiu5jH42"}},{"cell_type":"markdown","source":["1. Selected features"],"metadata":{"id":"y3pXJvEWjM-A"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","model_ridge = Ridge(alpha=100)\n","scores_select = cross_validate(model_ridge, X_train_select, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"kDNHLyGqe_1Z"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Cross-validation: R-squared\n","scores_select['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"4nFHNePBjZJC","executionInfo":{"status":"ok","timestamp":1639534137266,"user_tz":-60,"elapsed":13,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"31c4dd02-77c5-42ad-cce2-879b9e5654bb"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.65112807, 0.60841139, 0.65593906, 0.64538469, 0.6474118 ,\n","       0.66316019, 0.66337881, 0.65131412, 0.62789978, 0.62875737])"]},"metadata":{},"execution_count":67}]},{"cell_type":"code","source":["# Cross-validation: RMSE (negative)\n","scores_select['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"A3wa-R7fjkIS","executionInfo":{"status":"ok","timestamp":1639534137266,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"838565b8-3188-43eb-98b9-91282c90519d"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.56660922, -0.63120065, -0.56890592, -0.59166971, -0.59403042,\n","       -0.59012254, -0.57745015, -0.59202834, -0.61172992, -0.62121123])"]},"metadata":{},"execution_count":68}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores_select['test_r2'], 'g-')\n","ax2.plot(scores_select['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('RIDGE regression models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"id":"11JaGkofe_5n","executionInfo":{"status":"ok","timestamp":1639534137267,"user_tz":-60,"elapsed":10,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"2162c1e2-abb4-4901-ce70-581880165210"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["AS best performing in-sample model we'll chose model6!"],"metadata":{"id":"EpZ-zXOIkbvc"}},{"cell_type":"code","source":["# Intercept and coefficients\n","scores_select['estimator'][6].intercept_, scores_select['estimator'][6].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534137267,"user_tz":-60,"elapsed":10,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"d0ca6cb0-d110-491a-c6f5-f973edfb0bfa","id":"YebYZwYHkZ3_"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(-0.0012123553366365897,\n"," array([ 0.58517129, -0.2320018 , -0.8588067 , -0.78835302]))"]},"metadata":{},"execution_count":70}]},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_ridge_selected = scores_select['estimator'][6]"],"metadata":{"id":"OUIPTzIRkZ4A"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[3] = ['RIDGE (4 features)', scores_select['test_r2'][6], -scores_select['test_neg_root_mean_squared_error'][6], \"\"]"],"metadata":{"id":"crB2xFpykZ4A"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["2. All features:"],"metadata":{"id":"zx2KLajCk11S"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","scores_all = cross_validate(model_ridge, X_train, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"THZ0S5xYk11S"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["scores_all['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534137268,"user_tz":-60,"elapsed":9,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"cbd0f1b6-d107-4330-f64c-4b0b05f8d985","id":"Nub7RIsRk11S"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.6748906 , 0.62310505, 0.66951139, 0.66860664, 0.67069328,\n","       0.68235439, 0.67798425, 0.67707129, 0.64597592, 0.65283276])"]},"metadata":{},"execution_count":74}]},{"cell_type":"code","source":["scores_all['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534137268,"user_tz":-60,"elapsed":8,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"863bc852-0cd5-46a9-c4de-df7effb5a08e","id":"XyiZdncek11T"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.54697236, -0.61924509, -0.55757207, -0.571969  , -0.57408354,\n","       -0.57306242, -0.56478393, -0.56974254, -0.59668644, -0.60073059])"]},"metadata":{},"execution_count":75}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores_all['test_r2'], 'g-')\n","ax2.plot(scores_all['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('RIDGE regression models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1639534138138,"user_tz":-60,"elapsed":876,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"de54f4a7-efaf-4bdb-8554-ab99f3ab1157","id":"Mkljiguik11T"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["Here the model with the highest R-squared score is model 5. The Difference in RMSE between top performing models isn't so wide spread. we could make a compromise and select the model that optimizes the both score. In this case it would be model 0:\n","- Lowest RMSE\n","- Nearly the hugher R-squared"],"metadata":{"id":"91FFuZVkk11T"}},{"cell_type":"code","source":["# Intercept and coefficients\n","scores_all['estimator'][0].intercept_, scores_all['estimator'][0].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534138138,"user_tz":-60,"elapsed":14,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"20f9b1bc-2f66-4df1-feba-6acf738c9129","id":"HbB9QNyXk11T"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(0.001731834842365469,\n"," array([ 0.70488849,  0.12628465, -0.11303692,  0.10255655,  0.04191976,\n","        -0.22406762, -0.72973662, -0.66867968]))"]},"metadata":{},"execution_count":77}]},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_ridge_all = scores_all['estimator'][0]"],"metadata":{"id":"DS9tWbGCk11T"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[4] = ['RIDGE (all features)', scores_all['test_r2'][5], -scores_all['test_neg_root_mean_squared_error'][5], \"\"]"],"metadata":{"id":"YOR-xiRmk11T"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Out-of-the-sample evaluation**"],"metadata":{"id":"Se2W_oVKk11T"}},{"cell_type":"markdown","source":["1. Model with 4 features"],"metadata":{"id":"gPm-Lg8Zk11U"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_ridge_selected.predict(X_test_select)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534138139,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"d502a642-c1fa-4c8a-cdc1-1a7839a5c06b","id":"7I5P6Xt6k11U"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.6098316887555134"]},"metadata":{},"execution_count":80}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[3, 'RMSE (out-of-the-sample)'] = error"],"metadata":{"id":"blyLPx5Hk11U"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["2. Model with all features"],"metadata":{"id":"Zhqg9cjHk11U"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_ridge_all.predict(X_test)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534138140,"user_tz":-60,"elapsed":9,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"ddc6b65e-1f72-46e2-8dcf-eccbef9c711d","id":"qa5CbRo_k11U"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.5872467073646777"]},"metadata":{},"execution_count":82}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[4, 'RMSE (out-of-the-sample)'] = error"],"metadata":{"id":"mPczlQiHk11U"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Model comparison**\n","\n","As we can see from the table bellow, RIDGE regression model using only 4 features had a lower R-squared score in-sample, and in out-of-the-sample it performed slightly worse (higher RMSE). The RMSE for in-the-sample performance was nearly the same.\n","\n","All in all, RIDGE regression on all features was the best performing model."],"metadata":{"id":"8cKgDCL7k11U"}},{"cell_type":"code","source":["model_comparison.loc[3:5]"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":112},"executionInfo":{"status":"ok","timestamp":1639534138140,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"9549236e-973c-4baa-cf79-a74274e62928","id":"a4Z2n9V9k11U"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>R-squared (in-sample)</th>\n","      <th>RMSE (in-sample)</th>\n","      <th>RMSE (out-of-the-sample)</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>3</th>\n","      <td>RIDGE (4 features)</td>\n","      <td>0.663379</td>\n","      <td>0.577450</td>\n","      <td>0.609832</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>RIDGE (all features)</td>\n","      <td>0.682354</td>\n","      <td>0.573062</td>\n","      <td>0.587247</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                  Model  ...  RMSE (out-of-the-sample)\n","3    RIDGE (4 features)  ...                  0.609832\n","4  RIDGE (all features)  ...                  0.587247\n","\n","[2 rows x 4 columns]"]},"metadata":{},"execution_count":84}]},{"cell_type":"markdown","source":["#### 4.Lasso Regression"],"metadata":{"id":"i3rTvBPIbMIS"}},{"cell_type":"markdown","source":["Lasso regression similarly to the ridge regression adds a regularization term. The difference compared to the ridge regression is that  ℓ1  norm which defines the penalty term pulls many coeffcients to zero. Because LASSO regression is often used even as a predprocessing step for feature selection."],"metadata":{"id":"xJV2vYQOoKLy"}},{"cell_type":"markdown","source":["**In-sample evaluation**"],"metadata":{"id":"-IKiG20uomEz"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","lasso_model = LassoCV(cv=10)\n","scores = cross_validate(lasso_model, X_train, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"N3C9X297omEz"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Cross-validation: R-squared\n","scores['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534141643,"user_tz":-60,"elapsed":28,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"ca3a2e51-0c2d-4d90-b86f-03ebb5dc3f89","id":"gasJe8DHomEz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.67501994, 0.62350083, 0.66988005, 0.66907049, 0.67214237,\n","       0.6831441 , 0.67891142, 0.67630332, 0.64765244, 0.65396479])"]},"metadata":{},"execution_count":86}]},{"cell_type":"code","source":["# Cross-validation: RMSE (negative)\n","scores['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534141643,"user_tz":-60,"elapsed":23,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"34e887d2-7bba-468a-cfad-5ff45dea404d","id":"LF6wr54GomE0"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.54686355, -0.61891987, -0.55726099, -0.57156857, -0.57281904,\n","       -0.57234962, -0.56397026, -0.5704196 , -0.59527192, -0.59975037])"]},"metadata":{},"execution_count":87}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores['test_r2'], 'g-')\n","ax2.plot(scores['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('LASSO models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1639534141644,"user_tz":-60,"elapsed":20,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"3c3d35ff-63b0-49dc-901f-8960eb9c3d1e","id":"AF3K3BqVomE0"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["Again, the same compromise between the R-squared and RMSE score when choosing the best fit model. We can go with model 0, which gives us the best from both worlds"],"metadata":{"id":"2f499m0AomE0"}},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_lasso = scores['estimator'][0]"],"metadata":{"id":"5GDstCB9omE0"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Intercept and coefficients\n","scores['estimator'][0].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534141644,"user_tz":-60,"elapsed":16,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"37455633-f1cf-4329-e614-dbb6bbdf2b1d","id":"aLrrBrILomE0"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([ 0.69925813,  0.12023611, -0.10850869,  0.10465326,  0.03925168,\n","       -0.22371301, -0.79700011, -0.73528534])"]},"metadata":{},"execution_count":90}]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[5] = ['LASSO', scores['test_r2'][0], -scores['test_neg_root_mean_squared_error'][0], \"\"]"],"metadata":{"id":"fxBhVJJ5omE0"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Out-of-the-sample evaluation**"],"metadata":{"id":"DEh3of5momE0"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_lasso.predict(X_test)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534141645,"user_tz":-60,"elapsed":13,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"449c7909-1ef6-4d70-ccc3-5efc939a7706","id":"Z5ut33KQomE0"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.5862327657144357"]},"metadata":{},"execution_count":92}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[5, 'RMSE (out-of-the-sample)'] = error\n","model_comparison.loc[5]"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534141645,"user_tz":-60,"elapsed":11,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"2395e6b5-38aa-444e-ba33-8ff3cb839168","id":"G9cVBV0womE1"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Model                          LASSO\n","R-squared (in-sample)        0.67502\n","RMSE (in-sample)            0.546864\n","RMSE (out-of-the-sample)    0.586233\n","Name: 5, dtype: object"]},"metadata":{},"execution_count":93}]},{"cell_type":"markdown","source":["#### 5. Elastic Net Regression"],"metadata":{"id":"LUcvI3WZbX6V"}},{"cell_type":"markdown","source":["In the two previous sections, we we're modeling housing prices using RIDGE and LASSO regression models.\n","\n","In this section, we're going to work with `Elastic net regression`. It combines the best from both worlds; uses both the ℓ1  and  ℓ2  penalties.\n"],"metadata":{"id":"YDXz7cd4r9pG"}},{"cell_type":"markdown","source":["**In-sample evaluation**"],"metadata":{"id":"leES_jaer2By"}},{"cell_type":"code","source":["# Evaluate the models using crossvalidation\n","elastic_net_model = ElasticNetCV(cv=10)\n","scores = cross_validate(elastic_net_model, X_train, y_train, scoring=['r2','neg_root_mean_squared_error'], cv=10, return_estimator=True)"],"metadata":{"id":"a3IdsaR6r2Bz"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Cross-validation: R-squared\n","scores['test_r2']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534145284,"user_tz":-60,"elapsed":38,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"86550590-6cfb-436b-ce2b-2a1046d521a7","id":"kk1GN9Xpr2Bz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([0.67506751, 0.62350828, 0.66990187, 0.66905911, 0.67201128,\n","       0.6830973 , 0.67885122, 0.67646124, 0.64748707, 0.65387124])"]},"metadata":{},"execution_count":95}]},{"cell_type":"code","source":["# Cross-validation: RMSE (negative)\n","scores['test_neg_root_mean_squared_error']"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534145284,"user_tz":-60,"elapsed":33,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"7e066685-1913-4b78-a303-10d0d6ad0836","id":"78cXuZBUr2Bz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([-0.54682353, -0.61891375, -0.55724258, -0.5715784 , -0.57293355,\n","       -0.57239188, -0.56402313, -0.57028044, -0.5954116 , -0.59983143])"]},"metadata":{},"execution_count":96}]},{"cell_type":"code","source":["fig, ax1 = plt.subplots()\n","\n","ax2 = ax1.twinx()\n","ax1.plot(scores['test_r2'], 'g-')\n","ax2.plot(scores['test_neg_root_mean_squared_error'], 'r-')\n","\n","ax1.set_xlabel('Elastic Net models')\n","ax1.set_ylabel('R-squared', color='g')\n","ax2.set_ylabel('Negative RMSE', color='r')\n","\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1639534145284,"user_tz":-60,"elapsed":28,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"e9a85ead-fc7b-4b52-e8b9-5d4004537229","id":"8wYFN8aUr2Bz"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","source":["Again, the performance is nearly the same as for the previous models. We'll continue with model 0 (nearly the gighest score for both the R-squared and negative RMSE),"],"metadata":{"id":"YudgxSdQr2Bz"}},{"cell_type":"code","source":["# Model that we'll use for out of sample forecast\n","regression_elastic_net = scores['estimator'][0]"],"metadata":{"id":"b_dYqYg9r2Bz"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Intercept and coefficients\n","scores['estimator'][0].coef_"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534145285,"user_tz":-60,"elapsed":27,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"438ddcf8-818c-453a-ca2d-052fa2cf1338","id":"zEH5__7ur2Bz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([ 0.69991546,  0.1209169 , -0.1089752 ,  0.10426295,  0.03946997,\n","       -0.22370285, -0.78832464, -0.72668168])"]},"metadata":{},"execution_count":99}]},{"cell_type":"code","source":["# Saving performance results to dataframe\n","model_comparison.loc[6] = ['ELASTIC NET', scores['test_r2'][0], -scores['test_neg_root_mean_squared_error'][0], \"\"]"],"metadata":{"id":"yrA60_a2r2Bz"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Out-of-the-sample evaluation**"],"metadata":{"id":"TGnJoxSAr2Bz"}},{"cell_type":"code","source":["# Evaluation on test set\n","y_pred_select = regression_elastic_net.predict(X_test)\n","error = rmse(y_test, y_pred_select)\n","error"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534145285,"user_tz":-60,"elapsed":23,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"065958a1-2cd6-4133-e4fc-ff5aa11b1b85","id":"VgZ2xY9Vr2Bz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.5863200866800539"]},"metadata":{},"execution_count":101}]},{"cell_type":"code","source":["# Update model comparison table\n","model_comparison.loc[6, 'RMSE (out-of-the-sample)'] = error\n","model_comparison.loc[6]"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1639534145797,"user_tz":-60,"elapsed":531,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"db77338d-0f93-4d3f-8f86-0b54eace7f84","id":"jSZA3W6Zr2Bz"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Model                       ELASTIC NET\n","R-squared (in-sample)          0.675068\n","RMSE (in-sample)               0.546824\n","RMSE (out-of-the-sample)        0.58632\n","Name: 6, dtype: object"]},"metadata":{},"execution_count":102}]},{"cell_type":"markdown","source":["#### (b.2) Create a summary table (and/or a chart) that displays different model specifications and the corresponding in-sample and out-of-sample performance metrics."],"metadata":{"id":"MFF2PNI89Bxt"}},{"cell_type":"markdown","source":["Here is the complete overview of all models (one/best performing model from each regression group) that were evaluated in-the-sample and out-of-the-sample:"],"metadata":{"id":"_KTZ6V7NuUSH"}},{"cell_type":"code","source":["model_comparison"],"metadata":{"id":"I5pTfXDj9QSO","executionInfo":{"status":"ok","timestamp":1639534145797,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"colab":{"base_uri":"https://localhost:8080/","height":269},"outputId":"30100b35-4793-47bf-dcd1-050dcaa40fb0"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>R-squared (in-sample)</th>\n","      <th>RMSE (in-sample)</th>\n","      <th>RMSE (out-of-the-sample)</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>OLS (4 features)</td>\n","      <td>0.664642</td>\n","      <td>0.576366</td>\n","      <td>0.608098</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>OLS (all features)</td>\n","      <td>0.683187</td>\n","      <td>0.572311</td>\n","      <td>0.586047</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>PLS</td>\n","      <td>0.652611</td>\n","      <td>0.565404</td>\n","      <td>1.27231</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>RIDGE (4 features)</td>\n","      <td>0.663379</td>\n","      <td>0.577450</td>\n","      <td>0.609832</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>RIDGE (all features)</td>\n","      <td>0.682354</td>\n","      <td>0.573062</td>\n","      <td>0.587247</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>LASSO</td>\n","      <td>0.675020</td>\n","      <td>0.546864</td>\n","      <td>0.586233</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>ELASTIC NET</td>\n","      <td>0.675068</td>\n","      <td>0.546824</td>\n","      <td>0.58632</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                  Model  ...  RMSE (out-of-the-sample)\n","0      OLS (4 features)  ...                  0.608098\n","1    OLS (all features)  ...                  0.586047\n","2                   PLS  ...                   1.27231\n","3    RIDGE (4 features)  ...                  0.609832\n","4  RIDGE (all features)  ...                  0.587247\n","5                 LASSO  ...                  0.586233\n","6           ELASTIC NET  ...                   0.58632\n","\n","[7 rows x 4 columns]"]},"metadata":{},"execution_count":103}]},{"cell_type":"markdown","source":["For models that don't have embeded feature selection:\n","- OLS regression\n","- Ridge regression\n","we performed feature selection\n","\n","in table there are results for those models when fitted on all featured and when fitted on 4 most important features (determined using backward stepwise selection).\n","\n","For other models that have some sort of feature selection:\n","- PLS regression\n","- LASSO regression\n","- ELASTIC NET regression\n","\n","in the table we see the results for them as well."],"metadata":{"id":"lwKbDo21vELU"}},{"cell_type":"markdown","source":["#### (b.3) Draw a conclusion which features are most important and which models perform best, and discuss your results."],"metadata":{"id":"-mwDpG4T9BqT"}},{"cell_type":"markdown","source":["We already briefly discussed and compared model performance for every new group of models that we fittet.\n","\n","Now we'll just once again look at all the models and how how they performed both in-the-sample and out-of-the-sample."],"metadata":{"id":"T3JRHYzUrZQo"}},{"cell_type":"markdown","source":["First of all, we must notice that all of the models performed nearly the same. There is only a slight variation between them. The only outlier from that statement is PLS regression model, which had the worst RMSE on testing set.\n","\n","When looking the `R-squared` for `training set` the highest score have:\n","- OLS (all features)\n","- RIDGE (all features)\t\n","\n","Moving to the `RMSE` for the `trainig set` the lowest error that models have are models:\n","- LASSO\n","- ELASTIC NET\n","\n","And most importantly, how they all perform on `testing set`? The best four models, in term of the `lowest RMSE` are the previously mentioned four models and they have the error equal to the level of three decimal points (`0.586`):\n","- OLS (all features)\n","- RIDGE (all features)\t\n","- LASSO\n","- ELASTIC NET\n"," "],"metadata":{"id":"0Ssqn1Yjv4RV"}},{"cell_type":"markdown","source":["Now let's look at regression coefficients for our models:"],"metadata":{"id":"tecF4Tv5zNm0"}},{"cell_type":"code","source":["# Coefficients\n","columns = ['Model']\n","columns.extend([c for c in data_cleaned.columns[:-1]])\n","models = [regression_ols_all, regression_pls, regression_ridge_all, regression_lasso, regression_elastic_net]\n","regression_coef = pd.DataFrame([], columns=columns)\n","\n","for i, model in enumerate(models):\n","  row = [str(model)]\n","  row.extend(model.coef_)\n","  regression_coef.loc[i] = row\n","  "],"metadata":{"id":"E77DqWn9yL3b"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["regression_coef"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"Uu11k_bUzCAo","executionInfo":{"status":"ok","timestamp":1639534145798,"user_tz":-60,"elapsed":7,"user":{"displayName":"Nikola Krivacevic","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjddY-JeK_lvD3Syx0GbIgpRd6VoQ4ahFF3lzzCbQ=s64","userId":"06830379640663743667"}},"outputId":"4417efc5-d2c4-438b-e2f9-93b556817235"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Model</th>\n","      <th>MedInc</th>\n","      <th>HouseAge</th>\n","      <th>AveRooms</th>\n","      <th>AveBedrms</th>\n","      <th>Population</th>\n","      <th>AveOccup</th>\n","      <th>Latitude</th>\n","      <th>Longitude</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>LinearRegression()</td>\n","      <td>0.706126</td>\n","      <td>0.122501</td>\n","      <td>-0.109156</td>\n","      <td>0.106474</td>\n","      <td>0.04258</td>\n","      <td>-0.22211</td>\n","      <td>-0.800181</td>\n","      <td>-0.736939</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>PLSRegression(n_components=4)</td>\n","      <td>[0.8088212621974415]</td>\n","      <td>[0.18463030048918883]</td>\n","      <td>[-0.17242598627103706]</td>\n","      <td>[0.14307735200423388]</td>\n","      <td>[0.12823571902910466]</td>\n","      <td>[-0.23576524628490186]</td>\n","      <td>[-0.39613418795018457]</td>\n","      <td>[-0.3075105622168971]</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Ridge(alpha=100)</td>\n","      <td>0.704888</td>\n","      <td>0.126285</td>\n","      <td>-0.113037</td>\n","      <td>0.102557</td>\n","      <td>0.0419198</td>\n","      <td>-0.224068</td>\n","      <td>-0.729737</td>\n","      <td>-0.66868</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>LassoCV(cv=10)</td>\n","      <td>0.699258</td>\n","      <td>0.120236</td>\n","      <td>-0.108509</td>\n","      <td>0.104653</td>\n","      <td>0.0392517</td>\n","      <td>-0.223713</td>\n","      <td>-0.797</td>\n","      <td>-0.735285</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>ElasticNetCV(cv=10)</td>\n","      <td>0.699915</td>\n","      <td>0.120917</td>\n","      <td>-0.108975</td>\n","      <td>0.104263</td>\n","      <td>0.03947</td>\n","      <td>-0.223703</td>\n","      <td>-0.788325</td>\n","      <td>-0.726682</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                           Model  ...              Longitude\n","0             LinearRegression()  ...              -0.736939\n","1  PLSRegression(n_components=4)  ...  [-0.3075105622168971]\n","2               Ridge(alpha=100)  ...               -0.66868\n","3                 LassoCV(cv=10)  ...              -0.735285\n","4            ElasticNetCV(cv=10)  ...              -0.726682\n","\n","[5 rows x 9 columns]"]},"metadata":{},"execution_count":105}]},{"cell_type":"markdown","source":["Ignoring the PLS regression, which we previously saw that had the worst performance, we can see that the coefficients for all regression models are nearly the same.\n","\n","The least important feature is Population, coefficient for it is around 0.04. Median Income for district feature is one of the most important one. But the most important feature that higly affects the house price is location - Latitude and Longitude. Other coefficients, doesn't seem to be especially important. AveOccup with the average coeff. value of around 0.22 is somewhere important compared to the rest features."],"metadata":{"id":"xb7Mizeg1C9X"}}]}