ZCI_SS_Day3_Part1_Feature_selection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data Science Summer School - Split '16\n",
    "======================================"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Day 3 - part 1 - Feature selection\n",
    "--------------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Matija Piškorec\n",
    "\n",
    "v0.0.5\n",
    "\n",
    "`kernel: Python 2.7`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prerequisites\n",
    "--------------\n",
    "\n",
    "We will use [Seaborn](https://stanford.edu/~mwaskom/software/seaborn/) package for some of the plots and for visual style of Matplotlib's plots. If it is not already installed in your Anaconda distribution you can install it by `conda install seaborn` in Anaconda console. For more details see documentation for [`conda install`](http://conda.pydata.org/docs/commands/conda-install.html).\n",
    "\n",
    "We will use the \"Wisconsin Breast Cancer\" dataset which contains 569 samples and 30 features, with target variable being either 0 (malign) or 1 (benign). It is available locally in this notebook's folder, but in general it is also available in scikit-learn through [`sklearn.datasets.load_breast_cancer`](http://scikit-learn.org/dev/modules/generated/sklearn.datasets.load_breast_cancer.html#sklearn.datasets.load_breast_cancer) (version 0.17.1), and you can load it with:\n",
    "\n",
    "```python\n",
    "from sklearn.datasets import load_breast_cancer\n",
    "breast_cancer = load_breast_cancer()\n",
    "```\n",
    "\n",
    "And convert it to Pandas dataframe:\n",
    "\n",
    "```Python\n",
    "breast_cancer = pd.DataFrame(data=np.c_[breast_cancer['data'], breast_cancer['target']], \n",
    "                             columns=np.append(breast_cancer['feature_names'],'target'))\n",
    "```\n",
    "\n",
    "For other datasets available in scikit-learn see the documentation for [`sklearn.datasets`](http://scikit-learn.org/dev/modules/classes.html#module-sklearn.datasets)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import seaborn as sns; sns.set()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean radius</th>\n",
       "      <th>mean texture</th>\n",
       "      <th>mean perimeter</th>\n",
       "      <th>mean area</th>\n",
       "      <th>mean smoothness</th>\n",
       "      <th>mean compactness</th>\n",
       "      <th>mean concavity</th>\n",
       "      <th>mean concave points</th>\n",
       "      <th>mean symmetry</th>\n",
       "      <th>mean fractal dimension</th>\n",
       "      <th>...</th>\n",
       "      <th>worst texture</th>\n",
       "      <th>worst perimeter</th>\n",
       "      <th>worst area</th>\n",
       "      <th>worst smoothness</th>\n",
       "      <th>worst compactness</th>\n",
       "      <th>worst concavity</th>\n",
       "      <th>worst concave points</th>\n",
       "      <th>worst symmetry</th>\n",
       "      <th>worst fractal dimension</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>17.99</td>\n",
       "      <td>10.38</td>\n",
       "      <td>122.80</td>\n",
       "      <td>1001.0</td>\n",
       "      <td>0.11840</td>\n",
       "      <td>0.27760</td>\n",
       "      <td>0.3001</td>\n",
       "      <td>0.14710</td>\n",
       "      <td>0.2419</td>\n",
       "      <td>0.07871</td>\n",
       "      <td>...</td>\n",
       "      <td>17.33</td>\n",
       "      <td>184.60</td>\n",
       "      <td>2019.0</td>\n",
       "      <td>0.1622</td>\n",
       "      <td>0.6656</td>\n",
       "      <td>0.7119</td>\n",
       "      <td>0.2654</td>\n",
       "      <td>0.4601</td>\n",
       "      <td>0.11890</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20.57</td>\n",
       "      <td>17.77</td>\n",
       "      <td>132.90</td>\n",
       "      <td>1326.0</td>\n",
       "      <td>0.08474</td>\n",
       "      <td>0.07864</td>\n",
       "      <td>0.0869</td>\n",
       "      <td>0.07017</td>\n",
       "      <td>0.1812</td>\n",
       "      <td>0.05667</td>\n",
       "      <td>...</td>\n",
       "      <td>23.41</td>\n",
       "      <td>158.80</td>\n",
       "      <td>1956.0</td>\n",
       "      <td>0.1238</td>\n",
       "      <td>0.1866</td>\n",
       "      <td>0.2416</td>\n",
       "      <td>0.1860</td>\n",
       "      <td>0.2750</td>\n",
       "      <td>0.08902</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>19.69</td>\n",
       "      <td>21.25</td>\n",
       "      <td>130.00</td>\n",
       "      <td>1203.0</td>\n",
       "      <td>0.10960</td>\n",
       "      <td>0.15990</td>\n",
       "      <td>0.1974</td>\n",
       "      <td>0.12790</td>\n",
       "      <td>0.2069</td>\n",
       "      <td>0.05999</td>\n",
       "      <td>...</td>\n",
       "      <td>25.53</td>\n",
       "      <td>152.50</td>\n",
       "      <td>1709.0</td>\n",
       "      <td>0.1444</td>\n",
       "      <td>0.4245</td>\n",
       "      <td>0.4504</td>\n",
       "      <td>0.2430</td>\n",
       "      <td>0.3613</td>\n",
       "      <td>0.08758</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.42</td>\n",
       "      <td>20.38</td>\n",
       "      <td>77.58</td>\n",
       "      <td>386.1</td>\n",
       "      <td>0.14250</td>\n",
       "      <td>0.28390</td>\n",
       "      <td>0.2414</td>\n",
       "      <td>0.10520</td>\n",
       "      <td>0.2597</td>\n",
       "      <td>0.09744</td>\n",
       "      <td>...</td>\n",
       "      <td>26.50</td>\n",
       "      <td>98.87</td>\n",
       "      <td>567.7</td>\n",
       "      <td>0.2098</td>\n",
       "      <td>0.8663</td>\n",
       "      <td>0.6869</td>\n",
       "      <td>0.2575</td>\n",
       "      <td>0.6638</td>\n",
       "      <td>0.17300</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20.29</td>\n",
       "      <td>14.34</td>\n",
       "      <td>135.10</td>\n",
       "      <td>1297.0</td>\n",
       "      <td>0.10030</td>\n",
       "      <td>0.13280</td>\n",
       "      <td>0.1980</td>\n",
       "      <td>0.10430</td>\n",
       "      <td>0.1809</td>\n",
       "      <td>0.05883</td>\n",
       "      <td>...</td>\n",
       "      <td>16.67</td>\n",
       "      <td>152.20</td>\n",
       "      <td>1575.0</td>\n",
       "      <td>0.1374</td>\n",
       "      <td>0.2050</td>\n",
       "      <td>0.4000</td>\n",
       "      <td>0.1625</td>\n",
       "      <td>0.2364</td>\n",
       "      <td>0.07678</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean radius  mean texture  mean perimeter  mean area  mean smoothness  \\\n",
       "0        17.99         10.38          122.80     1001.0          0.11840   \n",
       "1        20.57         17.77          132.90     1326.0          0.08474   \n",
       "2        19.69         21.25          130.00     1203.0          0.10960   \n",
       "3        11.42         20.38           77.58      386.1          0.14250   \n",
       "4        20.29         14.34          135.10     1297.0          0.10030   \n",
       "\n",
       "   mean compactness  mean concavity  mean concave points  mean symmetry  \\\n",
       "0           0.27760          0.3001              0.14710         0.2419   \n",
       "1           0.07864          0.0869              0.07017         0.1812   \n",
       "2           0.15990          0.1974              0.12790         0.2069   \n",
       "3           0.28390          0.2414              0.10520         0.2597   \n",
       "4           0.13280          0.1980              0.10430         0.1809   \n",
       "\n",
       "   mean fractal dimension   ...    worst texture  worst perimeter  worst area  \\\n",
       "0                 0.07871   ...            17.33           184.60      2019.0   \n",
       "1                 0.05667   ...            23.41           158.80      1956.0   \n",
       "2                 0.05999   ...            25.53           152.50      1709.0   \n",
       "3                 0.09744   ...            26.50            98.87       567.7   \n",
       "4                 0.05883   ...            16.67           152.20      1575.0   \n",
       "\n",
       "   worst smoothness  worst compactness  worst concavity  worst concave points  \\\n",
       "0            0.1622             0.6656           0.7119                0.2654   \n",
       "1            0.1238             0.1866           0.2416                0.1860   \n",
       "2            0.1444             0.4245           0.4504                0.2430   \n",
       "3            0.2098             0.8663           0.6869                0.2575   \n",
       "4            0.1374             0.2050           0.4000                0.1625   \n",
       "\n",
       "   worst symmetry  worst fractal dimension  target  \n",
       "0          0.4601                  0.11890     0.0  \n",
       "1          0.2750                  0.08902     0.0  \n",
       "2          0.3613                  0.08758     0.0  \n",
       "3          0.6638                  0.17300     0.0  \n",
       "4          0.2364                  0.07678     0.0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 569 entries, 0 to 568\n",
      "Data columns (total 31 columns):\n",
      "mean radius                569 non-null float64\n",
      "mean texture               569 non-null float64\n",
      "mean perimeter             569 non-null float64\n",
      "mean area                  569 non-null float64\n",
      "mean smoothness            569 non-null float64\n",
      "mean compactness           569 non-null float64\n",
      "mean concavity             569 non-null float64\n",
      "mean concave points        569 non-null float64\n",
      "mean symmetry              569 non-null float64\n",
      "mean fractal dimension     569 non-null float64\n",
      "radius error               569 non-null float64\n",
      "texture error              569 non-null float64\n",
      "perimeter error            569 non-null float64\n",
      "area error                 569 non-null float64\n",
      "smoothness error           569 non-null float64\n",
      "compactness error          569 non-null float64\n",
      "concavity error            569 non-null float64\n",
      "concave points error       569 non-null float64\n",
      "symmetry error             569 non-null float64\n",
      "fractal dimension error    569 non-null float64\n",
      "worst radius               569 non-null float64\n",
      "worst texture              569 non-null float64\n",
      "worst perimeter            569 non-null float64\n",
      "worst area                 569 non-null float64\n",
      "worst smoothness           569 non-null float64\n",
      "worst compactness          569 non-null float64\n",
      "worst concavity            569 non-null float64\n",
      "worst concave points       569 non-null float64\n",
      "worst symmetry             569 non-null float64\n",
      "worst fractal dimension    569 non-null float64\n",
      "target                     569 non-null float64\n",
      "dtypes: float64(31)\n",
      "memory usage: 142.2 KB\n"
     ]
    }
   ],
   "source": [
    "breast_cancer.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although it is not strictly necessary in our case, we will follow a standard machine learning procedure and split our dataset into training and test set with [`sklearn.cross_validation.train_test_split`](http://scikit-learn.org/stable/modules/generated/sklearn.cross_validation.train_test_split.html), and perform all our analysis on the train set only."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples: 455\n",
      "Number of test examples: 114\n"
     ]
    }
   ],
   "source": [
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)\n",
    "print 'Number of training examples: ' + str(len(breast_cancer_train))\n",
    "print 'Number of test examples: ' + str(len(breast_cancer_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. Feature selection\n",
    "------------------\n",
    "\n",
    "The main goal of feature selection is to select a subset of most \"relevant\" features in order to 1) improve the predictive performance, 2) lower computational costs of predictive modeling and 3) improve the interpretability of models. It can be roughly divided in the following way:\n",
    "    \n",
    "1. *Model-independent methods* - That do not rely on a prediction model to evaluate feature relevance, but on various statistical tests such as Pearson's correlation coefficient, $\\chi^2$-statistics, mutual information, or ANOVA F-value. They are further dividided into *supervised* or *unsupervised* methods, depending on whether they evaluate the relevance using a target variable or not.\n",
    "    \n",
    "    \n",
    "2. *Model-dependant methods* - That rely on a specific predictive model to evaluate feature importances. If a predictive model provides feature relevance as a step in building the model, then these importances can be used directly. This is called a *transparent model* (or *embedded*) approach. Another approach is to evaluate feature relevance by their contribution to the predictive performance of the model by including or excluding them before building a model. This is called a *wrapper* approach. \n",
    "\n",
    "Note the difference between the *feature selection*, where we select a subset of existing features, and *dimensionality reduction*, where we project our data onto a different feature space. We will cover dimensionality reduction methods in the second part of these exercises.\n",
    "\n",
    "Scikit-learn [user guide](http://scikit-learn.org/stable/user_guide.html) has a section on feature selection [[1]](#scikit-feature-selection). Some of the following examples are from Chapter 4 of the book \"Python Machine Learning\" [[2]](#pyml-ch4) and Chapter 5 of the book \"Introduction to Machine Learning with Python\" [[3]](#intro-ml-py-ch5). Also, a good overview of feature selection is series of blog posts \"Selecting good features\" from [\"Diving into data\"](http://blog.datadive.net/) blog [[4]](#diving-into-data)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.1. Univariate feature selection\n",
    "-----------------------------------\n",
    "\n",
    "Probably the simplest unsupervised method for univariate feature selection is *variable variance* method, implemented in scikit-learn in [`sklearn.feature_selection.VarianceThreshold`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.VarianceThreshold.html#sklearn.feature_selection.VarianceThreshold), that excludes features which have variance below some threshold. Also, if we have two or more correlated features, we can keep just one of them, usually the one that is evaluated as the most relevant by other feature selection methods.\n",
    "\n",
    "In the following example we will perform a supervised univariate feature selection on the breast cancer dataset using ANOVA F-value as measure of feature's relationship with the (categorical) target variable, implemented in [`sklearn.feature_selection.f_classif`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.f_classif.html#sklearn.feature_selection.f_classif). To make our example more clear we will add noise features to our dataset. Statistically, they are unrelated to the target variable and we expect them to be easily recognized by simple univariate feature selection. We can use either [`sklearn.feature_selection.SelectKBest`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectKBest.html#sklearn.feature_selection.SelectKBest) or [`sklearn.feature_selection.SelectPercentile`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectPercentile.html#sklearn.feature_selection.SelectPercentile), depending whether we want to select certain number of features or certain proportion of features. As we know that we have 30 *real* features and 10 *noise* features, we will choose exactly 30 features to see whether we can discriminate between the two. For a list of all feature selection methods in scikit-learn see [`sklearn.feature_selection`](http://scikit-learn.org/stable/modules/classes.html#module-sklearn.feature_selection).\n",
    "\n",
    "For a similar example see [\"Univariate feature selection\"](http://scikit-learn.org/stable/auto_examples/feature_selection/plot_feature_selection.html) in scikit-learn guide."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import SelectKBest, f_classif"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean radius</th>\n",
       "      <th>mean texture</th>\n",
       "      <th>mean perimeter</th>\n",
       "      <th>mean area</th>\n",
       "      <th>mean smoothness</th>\n",
       "      <th>mean compactness</th>\n",
       "      <th>mean concavity</th>\n",
       "      <th>mean concave points</th>\n",
       "      <th>mean symmetry</th>\n",
       "      <th>mean fractal dimension</th>\n",
       "      <th>...</th>\n",
       "      <th>noise 0</th>\n",
       "      <th>noise 1</th>\n",
       "      <th>noise 2</th>\n",
       "      <th>noise 3</th>\n",
       "      <th>noise 4</th>\n",
       "      <th>noise 5</th>\n",
       "      <th>noise 6</th>\n",
       "      <th>noise 7</th>\n",
       "      <th>noise 8</th>\n",
       "      <th>noise 9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>17.99</td>\n",
       "      <td>10.38</td>\n",
       "      <td>122.80</td>\n",
       "      <td>1001.0</td>\n",
       "      <td>0.11840</td>\n",
       "      <td>0.27760</td>\n",
       "      <td>0.3001</td>\n",
       "      <td>0.14710</td>\n",
       "      <td>0.2419</td>\n",
       "      <td>0.07871</td>\n",
       "      <td>...</td>\n",
       "      <td>1.486066</td>\n",
       "      <td>2.259648</td>\n",
       "      <td>0.912368</td>\n",
       "      <td>-0.526343</td>\n",
       "      <td>-1.051469</td>\n",
       "      <td>0.005661</td>\n",
       "      <td>0.243024</td>\n",
       "      <td>1.672171</td>\n",
       "      <td>-0.014894</td>\n",
       "      <td>-0.399150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20.57</td>\n",
       "      <td>17.77</td>\n",
       "      <td>132.90</td>\n",
       "      <td>1326.0</td>\n",
       "      <td>0.08474</td>\n",
       "      <td>0.07864</td>\n",
       "      <td>0.0869</td>\n",
       "      <td>0.07017</td>\n",
       "      <td>0.1812</td>\n",
       "      <td>0.05667</td>\n",
       "      <td>...</td>\n",
       "      <td>0.635746</td>\n",
       "      <td>-0.069630</td>\n",
       "      <td>-0.457433</td>\n",
       "      <td>1.641570</td>\n",
       "      <td>-1.009987</td>\n",
       "      <td>-1.385897</td>\n",
       "      <td>0.715151</td>\n",
       "      <td>0.885629</td>\n",
       "      <td>1.868907</td>\n",
       "      <td>-0.177846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>19.69</td>\n",
       "      <td>21.25</td>\n",
       "      <td>130.00</td>\n",
       "      <td>1203.0</td>\n",
       "      <td>0.10960</td>\n",
       "      <td>0.15990</td>\n",
       "      <td>0.1974</td>\n",
       "      <td>0.12790</td>\n",
       "      <td>0.2069</td>\n",
       "      <td>0.05999</td>\n",
       "      <td>...</td>\n",
       "      <td>0.623077</td>\n",
       "      <td>0.398583</td>\n",
       "      <td>1.982409</td>\n",
       "      <td>-0.922211</td>\n",
       "      <td>-0.711093</td>\n",
       "      <td>0.630985</td>\n",
       "      <td>1.204588</td>\n",
       "      <td>0.698307</td>\n",
       "      <td>0.751816</td>\n",
       "      <td>-0.112009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.42</td>\n",
       "      <td>20.38</td>\n",
       "      <td>77.58</td>\n",
       "      <td>386.1</td>\n",
       "      <td>0.14250</td>\n",
       "      <td>0.28390</td>\n",
       "      <td>0.2414</td>\n",
       "      <td>0.10520</td>\n",
       "      <td>0.2597</td>\n",
       "      <td>0.09744</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.331564</td>\n",
       "      <td>-0.517568</td>\n",
       "      <td>0.393695</td>\n",
       "      <td>-1.377528</td>\n",
       "      <td>1.072365</td>\n",
       "      <td>-0.420694</td>\n",
       "      <td>-1.304702</td>\n",
       "      <td>0.589443</td>\n",
       "      <td>-1.503390</td>\n",
       "      <td>0.101499</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20.29</td>\n",
       "      <td>14.34</td>\n",
       "      <td>135.10</td>\n",
       "      <td>1297.0</td>\n",
       "      <td>0.10030</td>\n",
       "      <td>0.13280</td>\n",
       "      <td>0.1980</td>\n",
       "      <td>0.10430</td>\n",
       "      <td>0.1809</td>\n",
       "      <td>0.05883</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.215944</td>\n",
       "      <td>-0.634114</td>\n",
       "      <td>1.306986</td>\n",
       "      <td>0.209689</td>\n",
       "      <td>1.488484</td>\n",
       "      <td>0.664508</td>\n",
       "      <td>2.047157</td>\n",
       "      <td>0.429656</td>\n",
       "      <td>-1.275849</td>\n",
       "      <td>-0.398616</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean radius  mean texture  mean perimeter  mean area  mean smoothness  \\\n",
       "0        17.99         10.38          122.80     1001.0          0.11840   \n",
       "1        20.57         17.77          132.90     1326.0          0.08474   \n",
       "2        19.69         21.25          130.00     1203.0          0.10960   \n",
       "3        11.42         20.38           77.58      386.1          0.14250   \n",
       "4        20.29         14.34          135.10     1297.0          0.10030   \n",
       "\n",
       "   mean compactness  mean concavity  mean concave points  mean symmetry  \\\n",
       "0           0.27760          0.3001              0.14710         0.2419   \n",
       "1           0.07864          0.0869              0.07017         0.1812   \n",
       "2           0.15990          0.1974              0.12790         0.2069   \n",
       "3           0.28390          0.2414              0.10520         0.2597   \n",
       "4           0.13280          0.1980              0.10430         0.1809   \n",
       "\n",
       "   mean fractal dimension    ...      noise 0   noise 1   noise 2   noise 3  \\\n",
       "0                 0.07871    ...     1.486066  2.259648  0.912368 -0.526343   \n",
       "1                 0.05667    ...     0.635746 -0.069630 -0.457433  1.641570   \n",
       "2                 0.05999    ...     0.623077  0.398583  1.982409 -0.922211   \n",
       "3                 0.09744    ...    -0.331564 -0.517568  0.393695 -1.377528   \n",
       "4                 0.05883    ...    -0.215944 -0.634114  1.306986  0.209689   \n",
       "\n",
       "    noise 4   noise 5   noise 6   noise 7   noise 8   noise 9  \n",
       "0 -1.051469  0.005661  0.243024  1.672171 -0.014894 -0.399150  \n",
       "1 -1.009987 -1.385897  0.715151  0.885629  1.868907 -0.177846  \n",
       "2 -0.711093  0.630985  1.204588  0.698307  0.751816 -0.112009  \n",
       "3  1.072365 -0.420694 -1.304702  0.589443 -1.503390  0.101499  \n",
       "4  1.488484  0.664508  2.047157  0.429656 -1.275849 -0.398616  \n",
       "\n",
       "[5 rows x 41 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "\n",
    "N_noise_features = 10\n",
    "noise_features = pd.DataFrame(np.random.randn(breast_cancer.shape[0],N_noise_features),\n",
    "                              index=breast_cancer.index,\n",
    "                              columns=['noise ' + str(i) for i in range(N_noise_features)])\n",
    "\n",
    "breast_cancer_extended = pd.concat([breast_cancer, noise_features], axis=1)\n",
    "\n",
    "breast_cancer_extended.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer_extended, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "k_best = 30\n",
    "selector = SelectKBest(f_classif, k=k_best)\n",
    "selector.fit(X_train, y_train)\n",
    "scores = -np.log10(selector.pvalues_)\n",
    "indices = np.argsort(scores)[::-1] # return indices that would sort an array (in reverse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFdCAYAAAAufjWOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9965af0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.plot(range(X_train.shape[1]), scores[indices], 'o')\n",
    "\n",
    "plt.title('Univariate feature selection of best 30 features')\n",
    "plt.ylabel('log of the ANOVA p-value')\n",
    "plt.xlim([-1, X_train.shape[1]])\n",
    "plt.xticks(range(X_train.shape[1]), breast_cancer_train.drop('target',axis=1).columns.values[indices], rotation=90)\n",
    "plt.vlines(k_best-0.5, 0, np.max(scores), color='red', linestyle='--')\n",
    "plt.tight_layout()\n",
    "\n",
    "# Hack to color noise features in red\n",
    "[ i.set_color(\"red\") for i in plt.gca().get_xticklabels() if i.get_text()[:5] == 'noise' ]\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that some of our original features are actually ranked below some of the noise features. This is not surprising as the features in question are probably not very informative, either because they are very obscure ('mean fractal dimension') or they code an error of a feature. This is still not a guarantee that these features are useless, but it can serve as a simple first step in the feature selection process."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.1.** Instead of choosing specific number of features or a specific proportion of features to select, we can instead choose an upper limit on the proportion of *false discoveries* (that is, features for which we believe are related to the target variable when in fact they are not). Function [`sklearn.feature_selection.SelectFdr`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectFdr.html#sklearn.feature_selection.SelectFdr) selects features using a Benjamini-Hochberg procedure, given an upper bound on the expected false discovery rate `alpha`. Use this function to select features with FDR rate `alpha=0.05`. Hint: You can get a boolean mask for the selected features with `get_support()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import SelectFdr, f_classif"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "\n",
    "N_noise_features = 10\n",
    "noise_features = pd.DataFrame(np.random.randn(breast_cancer.shape[0],N_noise_features),\n",
    "                              index=breast_cancer.index,\n",
    "                              columns=['noise ' + str(i) for i in range(N_noise_features)])\n",
    "\n",
    "breast_cancer_extended = pd.concat([breast_cancer, noise_features], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer_extended, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']\n",
    "\n",
    "k_best = 30\n",
    "selector = SelectFdr(f_classif, alpha=0.05)\n",
    "selector.fit(X_train, y_train)\n",
    "scores = -np.log10(selector.pvalues_)\n",
    "indices = np.argsort(scores)[::-1] # return indices that would sort an array (in reverse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFcCAYAAADlIuYrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9ca6e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.plot(range(X_train.shape[1]), scores[indices], 'o')\n",
    "\n",
    "plt.title('Feature selection with false discovery rate (alpha = 0.5)')\n",
    "plt.ylabel('log of the ANOVA p-value')\n",
    "plt.xlim([-1, X_train.shape[1]])\n",
    "plt.xticks(range(X_train.shape[1]), breast_cancer_train.drop('target',axis=1).columns.values[indices], rotation=90)\n",
    "plt.vlines(np.sum(selector.get_support())-0.5, 0, np.max(scores), color='red', linestyle='--')\n",
    "plt.tight_layout()\n",
    "\n",
    "# Hack to color noise features in red\n",
    "[ i.set_color(\"red\") for i in plt.gca().get_xticklabels() if i.get_text()[:5] == 'noise' ]\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.2.** In the previous example we focused on a dataset with categorical target variable. Now, choose a dataset with a numerical target variable and perform univariate feature selection using a regression measure, for example [`sklearn.feature_selection.f_regression`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.f_regression.html#sklearn.feature_selection.f_regression). You can use one of the regression datasets available in [`sklearn.datasets`](http://scikit-learn.org/dev/modules/classes.html#module-sklearn.datasets), for example [`sklearn.datasets.load_diabetes`](http://scikit-learn.org/dev/modules/generated/sklearn.datasets.load_diabetes.html#sklearn.datasets.load_diabetes) or [`sklearn.datasets.load_boston`](http://scikit-learn.org/dev/modules/generated/sklearn.datasets.load_boston.html#sklearn.datasets.load_boston)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.2. Calculating feature relevances from a predictive model\n",
    "------------------------------------------------------\n",
    "\n",
    "In the example bellow we will calculate feature relevances directly using a random forest classifier, implemented in scikit-learn with [`sklearn.ensemble.RandomForestClassifier`](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier) (regression version is [`sklearn.ensemble.RandomForestRegressor`](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html#sklearn.ensemble.RandomForestRegressor)), which evaluates feature importance through the position in the tree where the feature is used as a split node. Features that are higher in the tree influence larger fraction of samples and can be considered as more important."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "forest = RandomForestClassifier(n_estimators=1000,random_state=0,n_jobs=-1) # change n_jobs for paralelism!\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']\n",
    "\n",
    "forest.fit(X_train, y_train)\n",
    "importances = forest.feature_importances_\n",
    "indices = np.argsort(importances)[::-1] # return indices that would sort an array (in reverse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFcCAYAAADlIuYrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9b24a10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.plot(range(X_train.shape[1]), importances[indices], 'o')\n",
    "\n",
    "plt.title('Feature importances from random forest')\n",
    "plt.ylabel('feature importances')\n",
    "plt.xlim([-1, X_train.shape[1]])\n",
    "plt.xticks(range(X_train.shape[1]), breast_cancer_train.drop('target',axis=1).columns.values[indices], rotation=90)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.3.** Each tree in a random forest ensemble evaluates feature relevances individually. In the above example we plotted only the mean estimate for all trees. Calculate and plot standard deviation of all estimates. Hint: Individual trees are stored in `forest.estimators_`, and each tree has its own `feature_importances_` attribute. You can use [`matplotlib.pyplot.errorbar`](http://matplotlib.org/api/pyplot_api.html?highlight=errorbar#matplotlib.pyplot.errorbar) function for plotting. Also, take a look at the scikit-learn example [\"Feature importances with forests of trees\"](http://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html#example-ensemble-plot-forest-importances-py).\n",
    "\n",
    "As each tree in the random forest ensemble evaluates feature importance individually, you can inspect the variance of the estimates:\n",
    "\n",
    "```Python\n",
    "importances_std = np.std([tree.feature_importances_ for tree in forest.estimators_],axis=0)\n",
    "```\n",
    "\n",
    "And plot it with the code above, using Matplotlib's `errorbar` instead of `plot`: \n",
    "\n",
    "```Python\n",
    "plt.errorbar(range(X_train.shape[1]), importances[indices], yerr=importances_std[indices], fmt='o')\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.4.** Use some other classifier for evaluation of feature importances, for example support vector machine classifier (SVC) implemented in [`sklearn.svm.SVC`](http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html). SVC calculates $k-1$ individual feature importances, where $k$ is the number of classes, and they are stored in `coef_` attribute (we have only two classes so we will get only one set of feature importances). Note: In case of SVC feature importances are available only for linear kernel. Performance note: [`sklearn.svm.SVC`](http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html) is based on [LIBSVM](http://www.csie.ntu.edu.tw/~cjlin/libsvm/) library. As an alternative, you can use [`sklearn.svm.LinearSVC`](http://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC) which is based on [LIBLINEAR](http://www.csie.ntu.edu.tw/~cjlin/liblinear/) library. It provides only linear kernel but scales better to large number of examples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.svm import SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "svc = SVC(kernel='linear')\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1).as_matrix()\n",
    "y_train = breast_cancer_train['target'].as_matrix()\n",
    "\n",
    "svc.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFcCAYAAADlIuYrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niWqCFLQlJCQkJCSqCf8PJsslMdHgjkIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9bdf6d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances_svc = svc.coef_.flatten() # as we have two classes we have just one set of feature importances\n",
    "indices_svc = np.argsort(importances_svc)[::-1] # return indices that would sort an array (in reverse)\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.plot(range(X_train.shape[1]), importances_svc[indices_svc], 'o')\n",
    "\n",
    "plt.title('Feature importances from a SVC classifier')\n",
    "plt.ylabel('feature importances')\n",
    "plt.xlim([-1, X_train.shape[1]])\n",
    "plt.xticks(range(X_train.shape[1]), breast_cancer.columns.values, rotation=90)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.3. Recursive feature elimination\n",
    "--------------------------------------------\n",
    "\n",
    "The procedure of the recursive feature elimination (also known as the *backward selection*) is the following. First, evaluate feature importances using the dataset with the full set of features, then iteratively eliminate the worst ranking feature and reevaluate. We repeat the procedure until only a desired number of features is left. Note that this is a simple greedy strategy that does not guarantee to yield best possible feature set of a given size.\n",
    "\n",
    "Similar procedure is *forward selection*, with the exception that it starts with empty set of features and iteratively adds features one by one until desired number of them is obtained. Unfortunately, it is not implemented in scikit-learn.\n",
    "\n",
    "We will use [`sklearn.feature_selection.RFE`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE) for recursive feature elimination. Following example is inspired by [\"Recursive feature elimination\"](http://scikit-learn.org/stable/auto_examples/feature_selection/plot_rfe_digits.html#example-feature-selection-plot-rfe-digits-py) in scikit-learn user guide. Recursive feature elimination works on any predictive model that assigns feature importances. In the example below we will use random forest classifier implemented in [`sklearn.ensemble.RandomForestClassifier`](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "rfe_rf = RFE(RandomForestClassifier(n_estimators=100,random_state=0,n_jobs=-1), n_features_to_select=10, step=1)\n",
    "rfe_rf.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected features with RFE (using random forest):\n",
      " - mean texture\n",
      " - mean area\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - worst radius\n",
      " - worst texture\n",
      " - worst perimeter\n",
      " - worst area\n",
      " - worst concavity\n",
      " - worst concave points\n"
     ]
    }
   ],
   "source": [
    "print \"Selected features with RFE (using random forest):\"\n",
    "for i in list(breast_cancer_train.columns.values[:-1][rfe_rf.get_support()]): print ' - ' + str(i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.5.** Instead of random forest, use logistic regression implemented in [`sklearn.linear_model.LogisticRegression`](http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression) and support vector machine classifier implemented in [`sklearn.svm.SVC`](http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC) for recursive feature elimination of 10 most informative features. Compare the features you obtained with the ones above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rfe_svc = RFE(SVC(kernel='linear'), n_features_to_select=10, step=1)\n",
    "rfe_svc.fit(X_train, y_train);\n",
    "\n",
    "rfe_lr = RFE(LogisticRegression(), n_features_to_select=10, step=1)\n",
    "rfe_lr.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Selected features with RFE (using SVC):\n",
      " - mean radius\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - texture error\n",
      " - worst radius\n",
      " - worst smoothness\n",
      " - worst compactness\n",
      " - worst concavity\n",
      " - worst concave points\n",
      " - worst symmetry\n",
      "\n",
      "Selected features with RFE (using logistic regression):\n",
      " - mean radius\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - radius error\n",
      " - perimeter error\n",
      " - worst radius\n",
      " - worst perimeter\n",
      " - worst compactness\n",
      " - worst concavity\n",
      " - worst concave points\n"
     ]
    }
   ],
   "source": [
    "print \"\\nSelected features with RFE (using SVC):\"\n",
    "for i in list(breast_cancer.columns.values[:-1][rfe_svc.get_support()]): print ' - ' + str(i)\n",
    "    \n",
    "print \"\\nSelected features with RFE (using logistic regression):\"\n",
    "for i in list(breast_cancer.columns.values[:-1][rfe_lr.get_support()]): print ' - ' + str(i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1.4. Recursive feature elimination with cross-validation\n",
    "--------------------------------------------\n",
    "\n",
    "While performing recursive feature elimination we had to define a number or a proportion of features we want to select in advance. But how many features is optimal? It would be more convenient if an algorithm could choose it for us. One of the ways to do it is with cross-validation. Feature importance is still used to decide which features to eliminate, but the optimal number of features is selected with cross-validation. For this we will use [`sklearn.feature_selection.RFECV`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.RFECV.html#sklearn.feature_selection.RFECV) with stratified 2-fold cross-validation implemented in [`sklearn.cross_validation.StratifiedKFold`](http://scikit-learn.org/stable/modules/generated/sklearn.cross_validation.StratifiedKFold.html#sklearn.cross_validation.StratifiedKFold). For all available types of cross validation in scikit-learn see [`module-sklearn.cross_validation`](http://scikit-learn.org/stable/modules/classes.html#module-sklearn.cross_validation). For a similar example see [\"Recursive feature elimination with cross-validation\"](http://scikit-learn.org/stable/auto_examples/feature_selection/plot_rfe_with_cross_validation.html#example-feature-selection-plot-rfe-with-cross-validation-py) in scikit-learn user guide."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import StratifiedKFold\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "rfecv_rf = RFECV(RandomForestClassifier(n_estimators=100,random_state=0,n_jobs=-1), \n",
    "                 step=1, \n",
    "                 cv=StratifiedKFold(y_train, 2),\n",
    "                 scoring='accuracy')\n",
    "rfecv_rf.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected features with RFECV (using random forest):\n",
      " - mean radius\n",
      " - mean texture\n",
      " - mean perimeter\n",
      " - mean area\n",
      " - mean smoothness\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - radius error\n",
      " - area error\n",
      " - compactness error\n",
      " - concavity error\n",
      " - fractal dimension error\n",
      " - worst radius\n",
      " - worst texture\n",
      " - worst perimeter\n",
      " - worst area\n",
      " - worst smoothness\n",
      " - worst compactness\n",
      " - worst concavity\n",
      " - worst concave points\n",
      " - worst symmetry\n",
      " - worst fractal dimension\n"
     ]
    }
   ],
   "source": [
    "print \"Selected features with RFECV (using random forest):\"\n",
    "for i in list(breast_cancer.columns.values[:-1][rfecv_rf.get_support()]): print ' - ' + str(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcQAAAFRCAYAAADq9N3vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQggBwP8BgEE+u2KqOMwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9d444d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(7,5))\n",
    "\n",
    "plt.plot(range(1, len(rfecv_rf.grid_scores_) + 1), rfecv_rf.grid_scores_, '-o')\n",
    "plt.xlabel('Number of features selected')\n",
    "plt.ylabel('cross-validation score')\n",
    "plt.title('random forests')\n",
    "plt.vlines(rfecv_rf.n_features_, \n",
    "           np.min(rfecv_rf.grid_scores_), \n",
    "           rfecv_rf.grid_scores_[rfecv_rf.n_features_-1], \n",
    "           color='red', linestyle='--')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1.6.** Instead of random forest and stratified 2-fold cross-validation use some other classifier and type of cross-validation. For example, for classifier you could use support vector machine classifier (SVC) implemented in [`sklearn.svm.SVC`](http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC) or logistic regression implemented in [`sklearn.linear_model.LogisticRegression`](http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression). For cross-validation you can use leave-one-out cross-validation implemented in [`sklearn.cross_validation.LeaveOneOut`](http://scikit-learn.org/stable/modules/generated/sklearn.cross_validation.LeaveOneOut.html#sklearn.cross_validation.LeaveOneOut) or shuffle split crossvalidation implemented in [`sklearn.cross_validation.ShuffleSplit`](http://scikit-learn.org/stable/modules/generated/sklearn.cross_validation.ShuffleSplit.html#sklearn.cross_validation.ShuffleSplit). See [`module-sklearn.cross_validation`](http://scikit-learn.org/stable/modules/classes.html#module-sklearn.cross_validation) for all available cross-validation types in scikit. Hint: You can use [`matplotlib.pyplot.subplot`](http://matplotlib.org/api/pyplot_api.html?highlight=subplot#matplotlib.pyplot.subplot) for plotting multiple figures."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import StratifiedKFold\n",
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breast_cancer = pd.read_csv('breast_cancer.csv',index_col=0)\n",
    "breast_cancer_train, breast_cancer_test = train_test_split(breast_cancer, test_size=0.20, random_state=42)\n",
    "\n",
    "X_train = breast_cancer_train.drop('target',axis=1)\n",
    "y_train = breast_cancer_train['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rfecv_svc = RFECV(SVC(kernel='linear'), \n",
    "                 step=1, \n",
    "                 cv=StratifiedKFold(y_train, 2),\n",
    "                 scoring='accuracy')\n",
    "rfecv_svc.fit(X_train, y_train);\n",
    "\n",
    "rfecv_lr = RFECV(LogisticRegression(), \n",
    "                 step=1, \n",
    "                 cv=StratifiedKFold(y_train, 2),\n",
    "                 scoring='accuracy')\n",
    "rfecv_lr.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Selected features with RFECV (using SVC):\n",
      " - mean radius\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - texture error\n",
      " - worst radius\n",
      " - worst smoothness\n",
      " - worst compactness\n",
      " - worst concavity\n",
      " - worst concave points\n",
      " - worst symmetry\n",
      "\n",
      "Selected features with RFECV (using logistic regression):\n",
      " - mean radius\n",
      " - mean texture\n",
      " - mean perimeter\n",
      " - mean smoothness\n",
      " - mean compactness\n",
      " - mean concavity\n",
      " - mean concave points\n",
      " - mean symmetry\n",
      " - mean fractal dimension\n",
      " - radius error\n",
      " - texture error\n",
      " - perimeter error\n",
      " - area error\n",
      " - concavity error\n",
      " - concave points error\n",
      " - symmetry error\n",
      " - worst radius\n",
      " - worst texture\n",
      " - worst perimeter\n",
      " - worst smoothness\n",
      " - worst compactness\n",
      " - worst concavity\n",
      " - worst concave points\n",
      " - worst symmetry\n",
      " - worst fractal dimension\n"
     ]
    }
   ],
   "source": [
    "print \"\\nSelected features with RFECV (using SVC):\"\n",
    "for i in list(breast_cancer.columns.values[:-1][rfecv_svc.get_support()]): print ' - ' + str(i)\n",
    "    \n",
    "print \"\\nSelected features with RFECV (using logistic regression):\"\n",
    "for i in list(breast_cancer.columns.values[:-1][rfecv_lr.get_support()]): print ' - ' + str(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4IAAAFRCAYAAAAhPBPJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XlclOXeP/DPLDCAw7CJC4iCCoqguJw0NEtNi9LMNfWU\n6aN26uloi51OVmaW+vC0nE7HjrZbT+XROv5yrzSXloMIioKxiKIiq8i+zcDAzP37A5lEZhgGmI37\n8369er3ivu9r7u+FMBffua/vdUkEQRBAREREREREoiG1dwBERERERERkW0wEiYiIiIiIRIaJIBER\nERERkcgwESQiIiIiIhIZJoJEREREREQiw0SQiIiIiIhIZJgIEjmQ5ORkPProo3jwwQfxwAMP4E9/\n+hOysrLw6KOP4qOPPmp1/bZt2/Dkk08CAPR6PT777DPMnTsXs2fPxowZM/D2229Dq9XauhtEROTk\nEhMT8cADD3S4/bFjx7Bp06Y2r/n555+xefPmdl/vTFJTU/H000/bOwyiNkm4jyCRY9Bqtbjzzjvx\n+eefY+jQoQCAffv24d1338Xzzz+Pd999F4cOHWrR5r777sO6desQHR2NV155BdXV1di4cSOUSiXq\n6urw3HPPQalU4o033rBHl4iIyEklJiZiw4YN2L9/v9Xu8c9//hMVFRVYu3at1e5BRKbJ7R0AETWp\nq6tDTU0NampqDMdmzpwJT09P3HHHHfjf//1fJCUlYcyYMQCaBmkAiI6ORm5uLg4cOIC4uDh4eHgA\nANzc3PD666/j7Nmztu8MERF1GzU1NXjttddw/vx5SCQSTJw4Ec899xykUil+/vlnvP3225DL5Rg6\ndChOnDiBHTt2ICEhAYcOHcIHH3yAw4cP44MPPoBUKoVMJsPzzz8PV1dX7Ny5E3q9HkqlEgMGDDBc\nX1JSgldffRWXL1+GTCbDggULsHjx4hYx7d69G7t27YJGo4Gnpyf+7//+D//+97+xY8cOAIC3tzfW\nrl2LgQMHoqysDC+99BJyc3Ph7e0NPz8/hIWFYeXKlYiMjMTUqVORmZmJt99+G25ubvif//kfVFRU\nQK/XY/HixZgzZw7UajVefPFF5OTkQCKRIDIyEq+//rrJ4zcn0m19/0aMGIE//elPiIuLQ3FxMRYv\nXowlS5bY45+ZRIiJIJGDUKlU+Mtf/oIVK1bA398fo0aNwrhx4zB9+nS4uLhg/vz52LVrlyER/Oab\nb/Dwww8DADIyMhAaGmpIApv5+flh6tSpNu8LERF1Hxs2bICPjw/279+PhoYGPPHEE/j0008xf/58\n/PWvf8WXX36JsLAw7NmzB3v27GnV/q233sLf/vY3jBgxAidOnEBiYiKefPJJLFy4EBUVFXjmmWew\ne/duw/Xr169HSEgItmzZgpqaGixatAiTJk1CUFBQi9fNysrC8ePH4eHhgVOnTmHv3r3YsWMHFAoF\n4uLisGrVKhw8eBAbN25EaGgoPvjgAxQXF2POnDkICwsDADQ2NuLuu+/Gu+++C51OhwcffBBvvfUW\nwsPDUVNTgwULFmDw4MG4cuUK1Go1du/eDb1ej/Xr1yM3Nxdnzpwxerw937/HHnsMWq0Wvr6+2LFj\nB9LS0rBo0SIsWrQIrq6uVviXJGqJNYJEDmTp0qU4ceIE1q5di169euHjjz/G7NmzDYPRsWPHoFar\nUVFRgbi4OMyaNQsAIJVKodfr7Rw9ERF1R7/++iseeeQRAICLiwsWLVqEX375BadPn0ZoaKghqZo1\naxaUSmWr9tOnT8eTTz6JtWvXoqKiAitWrGjzfvHx8ViwYAEAQKlUYv/+/a2SQAAYMmSI4QPQn376\nCTk5OVi4cCFmzZqFt956C1VVVaisrMQvv/yChx56CADg7++Pe++9t8XrNH/Amp2djZycHLz00kuY\nNWsWHnnkEdTX1yM9PR1jxoxBVlYWFi9ejI8++ghLlixBUFCQyePt+f41u/vuuwEAERERaGhogEaj\nafP7Q9RV+ESQyEGcOXMGZ8+exfLly3HXXXfhrrvuwurVqzFjxgycOHEC99xzD8aPH4+DBw9CrVbj\n3nvvNQy4w4cPx6VLl6BWq1s8FSwqKsK6devw3nvv8dNFIiLqkFs/aNTr9WhsbIRcLm91TiKRtGr/\nzDPPYN68eYiLi8Pu3bvx8ccft3gCeCu5vOWfp7m5ufDx8WmVZN483un1ejz44IN47rnnDMeKiorg\n5eXV6vVkMpnR19HpdFCpVC1iKy0thaenJ1xdXXH48GEkJibi5MmTWLJkCdatW4d77rkHhw4dwqlT\np1oc9/b2bhHbzZq/f80UCkWL81y+g2yFTwSJHISvry8++OADnDlzxnCsqKgIdXV1hk9bFy1ahH37\n9mHv3r2GaaEA0Lt3bzzwwAN46aWXDDWGzTUJvr6+TAKJiKjD7rjjDmzfvh1A08JmX3/9NSZMmIBR\no0bh6tWruHDhAgDg0KFDqK6ubpEM6nQ6TJkyBWq1GgsWLDDU/jU0NEAmk6GhoaHV/caPH49vv/0W\nAFBdXY2lS5ciJyenzRgnTJiAgwcPori4GACwfft2LF26FAAwadIk7Nq1CwBQXl6OH3/80WjCGhIS\nAoVCgX379gEACgsLMWPGDKSlpWHHjh1Ys2YNJkyYgOeeew4TJ07EhQsXsGPHDrz44outjrfn+2cM\nk0CyJT4RJHIQwcHB2LJlC9555x0UFRVBoVBAqVRiw4YNCA4OBgCMHTsWFRUV8PHxQWhoaIv269ev\nx5YtW7Bo0SLI5XJotVpMnToVq1atskNviIiou1i7di02bNiABx54AA0NDbjzzjvxxBNPQC6X4+23\n38Zf//pXSKVSREZGQiaTwc3NzdBWJpPh5ZdfxnPPPQcXFxdIpVLExsbCxcUF0dHRWLVqFVxcXBAR\nEWFo88orr2D9+vWYOXMmBEHAE088gWHDhrUZ4x133IEVK1Zg2bJlkEqlUCqV+Oc//wkAWLNmDdau\nXYuZM2fC29sbgYGBcHd3B9DyCaaLiwu2bt2KjRs34pNPPoFOp8Ozzz6LUaNGYejQoTh16hTuv/9+\nuLu7IzAwEEuWLIFcLkdiYmKr4xkZGSa/fxMnTsQTTzzR6v7GviayJm4fQUREREQWq6mpwfvvv4+n\nnnoKCoUC6enpePzxx/Hrr7/aO7QW/vWvfyEiIgJRUVHQarV4+OGH8dRTT2HixIn2Do3Irqz6RFAQ\nBKxfvx6ZmZlwdXXFpk2bWhTQ7tmzB9u2bYNKpcKsWbMwb948AMBHH32EY8eOoaGhAX/84x8xd+5c\na4ZJRERERBZSKpVwcXHB3LlzIZfL4eLign/84x/2DquVwYMH4/XXXzfU5sXExDAJJIKVnwj++OOP\nOHbsGGJjY5GSkoIPP/wQW7duBdA0R3vu3LnYu3cvlEolli5ditjYWOTl5eGzzz7D+++/D7VajW3b\ntmHlypXWCpGIiIiIiEh0rPpEMCkpyfCJS1RUFFJTUw3ncnNzER4eDk9PTwBNqx4mJyfj/PnzCAsL\nw5NPPona2lr89a9/tWaIREREREREomPVRLCmpsaQ6AEwLDMslUoRHByMrKwslJWVwd3dHfHx8QgJ\nCUF5eTkKCgrw4YcfIjc3F//93/+NH374wZphEhERERERiYpVE0GlUona2lrD181JIACoVCqsWbMG\nq1atgre3NyIiIuDj4wNvb28MGjQIcrncsIxvWVkZfH19Td5HEASuskRERHSLxkYdysvV9g7Dbnx8\nPETbfzH3HRB3/9l3cfYdAPz9Pc1fdBOrJoKjR4/G8ePHERMTg+TkZMNeaEDTvjJpaWnYvn07tFot\nli9fjtWrV0MqleLLL7/E0qVLDXuo+fj4tHkfiUSC4uJqa3bFYfn7e4q274C4+8++i7PvgLj7b+kg\nJ3Zyucz8Rd2YmPsv5r4D4u4/+07tZdVEcNq0aYiLi8PChQsBALGxsThw4AA0Gg3mz58PAJg9ezYU\nCgWWLVsGb29vTJo0CadPn8a8efMgCAJeffVVPu0jIiIiIiLqQt1mH0Exfzou1r4D4u4/+y7OvgPi\n7j+fCFpOrD8rAH9XxNp3QNz9Z9/F2XfA8jFSaqU4iIiIiIiIyEExESQiIiIiIhIZJoJEREREREQi\nw0SQiIiIiIhIZJgIEhERERERiYxVt48gspaE9CIcjM9GQakaAX4emB4djHHDets7LCIiIiIip8BE\nkJxOQnoRPtyXZvg6r7jW8DWTQSIiIiIi8zg1lJzOwfhsE8ev2jQOIiIiIiJnxUSQnE5Bidro8cLS\nWhtHQkRERETknDg1lJxOQE8P5BW3Tvq8lQo7REO21JHaUEObEjUCejpOm07dQ4S1sTf3fe9bM+0d\nDhERkdOTrV+/fr29g+gKarXW3iHYRY8eCtH13V0hR9KF4lbHNfWNuF6uQfgAb7jIZXaIzLbE9m/f\nXBtapW6AIABV6gYkZRajj68H+vkrzbeB47Tp9D3a2f/u4ta+//HeofYOyamI6X3iVmJ7n7yZmPsO\niLv/7Ls4+w409d8SfCJITieotycAwEMhR32DDn39emB8ZB8kZhQhPu0azueUY9n0cEQE+9o5UupK\npmpDvzl2Eeq6BqPn9p9wzDZdeY+D8Ve7/VNBU//2RERE1HFMBMnpZGSXAQAWTBmMOVOHoLi4GgAw\n9Q/9cDD+KvbHZeNvO5Nx95h+mDdpEBQu3f/pYHdXXl2P/BLjNaDlNVp8efiCZa/noG06cg8x1Maa\nqgsmIiKijmMiSE4nPbscABAe7NPiuFwmxYN3hGDEID98ciAdR5PykHalDCtmDMPAAJU9QqVOqFZr\nkZRZjIT0IlzIrYBg4jo/lQLzJw82eu6b41koq6p3uDZdeQ9/b3ej13cnpuqCiYiIAMddQ8Dm6xRY\nWEfPRJCcik6vR2ZuOXr5uCNsym2AVAKc+q3FNSF9VXh16W34fz9fxo+nc/E/XyZhVKgfrpVpUFhq\n3V9ER+Xwb1w32ky7LQgyqQQJ6deRnl0Gnb4p/Qvt54XePu74z2/XWr3GvEmDMTbc+H0EAS32nHSU\nNl15j+IKDY4m5WHK6EBIJBKjbZ1d5EA/JoJERGSUqf2l6xsaMWZIL6NtkjKv4/PvM63axhb3MNbG\nEhJBEEx90O5UmqcHio2/v6eo+n4pvxKbvkzCpJEBeObleZBJJSi+JRG8WcbVcry/5zfUaBpbnVt6\n3xD8wcQv1WkTv1SPz4xwmGSwvf/2t75BNlsxI9xo0pGYUYRPDmS0+/qubtNsQB9PjAvvjbHhveCr\ncjP05WD8VRSW1qKvXw9Mjx7QzmTT8dp0xT2G9vdGfNo11NY1IiLYB/91f7jhe9VdlFfX49VtiVDX\nNcLf2w0llXXYw1VDLSKmMeJWYhsjbybmvgPi7r/Y+r7u0wR+WHiT/X97sN3XMhF0cmL7Zd9/Ihu7\nf7mMJ2dF4p6Hp5hNBAFg7ScJKDBRX2apfv5KvL58bJe8Vme159++oVGPlz46idKqOhtF1XmqHq5Y\n8/Bo9PH1MHmN2H7ub3Vz/8ur6/H59+fx2+VSeCjkePieMNw+rHe3eDqoFwT8bWcyMq6W449TQzH1\nD0Hw9/e0d1hOh78r4uy/mPsOiLv/Yuv7ijeOQ28inRkV2tPo8bMXS0y+Xle1scU9jLWxJBHk1FBy\nKhnZZZAAGDrAx+y1za6Vml5owtJfxAInWJhDp9fj/NUKJGQU4UxmMdT1rZ+GNosIbv19TLtRg9ne\n67u6Ta2moc0kkFry8VTgmfkj8HNKAb4+moWP96fj7IViLL53CDw9XO0dXqf8kJCDjKvliBrkh7vH\n9LN3OERE5IBM1ZH381di1dwRRtuYeorYlW1scY+22rQHE0FyGvUNOmTlV6J/b08o3V3a3a4r3yBk\nUgnyimu6bN+2rtpU/LbwXsjKq0RCRhFOn7+OanXTFgQ+ngpIpRLUaFpvSdDPX4nnFo5qdbytNyFj\n13d1m75+PYxeT6ZJJBJMGhmIYQN88MnBDJzOLMbFvEosvW8o6rQ6p6x1vVxQhd2/XIaX0hXLpod3\niyecRETU9abdFoTPvjvf6vj06AEm20yPDjZaNtOVbWxxj7batAc3lHdyYto483xOOeJ+u4boyD6I\nCPGF+0dbIZVIoP7Tk22283BzQVJm6w3oF00NNZnQmWqj1wv49VwBXOQyDAxUdeqP067cVPxoUh6O\nn81HdmE13FxkmDC8LxZMCcXCu0Php3KzqP9d+f3q6jbNxPRzb4yp/vdwd8GEyL5QuMhw7nIp4tOK\nkJRZbNHPmCPQ1Dfina+TUatpwMo5w1vEaulmuSTe8REQ93uFmPsOiLv/Yuv7uawSnM+pgKqHKxp1\negT2VGLR1NA2P/Ts569EH18PFJVpUFvXYJU2triHsTYL7xli8tpb8YkgOY3mbSOG3ZhqWJaU2lQv\nZGYefPMvjyULc5hqo3CV4fPvMvDN8SykZJVg+Yxw9PTq2PL9pjbJ/uJQJs5ebJ0gAcBvl0uNHlfX\nN+KOEX0xLrw3hg7whkwqNdsXU/3vyu9XV7ch86RSCe67fQAiB/ph4/+dRoNO3+oaR9+E/qvDmbhe\nocH9tw/AsGBfe4djFYIgYP369cjMzISrqys2bdqEoKAgw/k9e/Zg27ZtUKlUmDVrFubNm2c4V1pa\nirlz5+Kzzz5DSEiIPcInInIItXUN+CExF0p3F8T+6Xb07+fT7vrIccN6WzwWWtrGFve4uY2ldfRM\nBMlpZGSXQy6TILSft8Vtu/IX8fUV4/DFD5k4c6EY6z5NxB+nhmHC8D7tfjqormvE2YvFJudza+ob\nkZhx3aJYJRIJlt0fbvK8o79xUdcL6qWETt86CQQcexP6+NRriE8rQkhfFWZN7L5JzpEjR6DVarFz\n506kpKQgNjYWW7duBQCUl5dj8+bN2Lt3L5RKJZYuXYrx48cjICAAjY2NePXVV+Hm1r1WhyUi6ogf\nEnKgqW/EQ5MHI3D8SKPbipFpTATJKdRoGpBTVI0h/b2hcJHZNRaVhyv+PDsSJ1KvYfuPF7Dtuwwk\nZ5Xg0ZghUJlYnKNeq0PKpRIkpBfht8ulaNSZXqy3r58Hnl9kvK7urR1nUWhk8ZsA1tWREQE9exj9\nwKGnl2MmEdfL1fjicCbcXGV4fOYwyGVS842cVFJSEiZOnAgAiIqKQmpqquFcbm4uwsPD4enZ9Mnu\n8OHDkZycjICAALzxxhtYtGgRPvzwQ7vETUTkKKpqtThyOg9eSldMHh1o73CcEhNBcgrnr5ZDABBu\nwWqh1iSRSDBheF8MCfLGJwczcOZCMbLyKjB+eF+kXi5FQYkaff08EDnQFxU1WiRfLEF9gw4AEOjf\nA2PDe0PhIsXOo1mtXnvmhBB4K43XQc2cEGJxETGJl6kC8uIKDb5PuIp7b+sPqdQxFmFp1OmbNs3V\n6vDYA8PQy6d7rxxbU1NjSPQAQC6XQ6/XQyqVIjg4GFlZWSgrK4O7uzvi4+MREhKC3bt3w8/PDxMm\nTMAHH3xgx+iJiOzvu5NXUd+gw7xJg+z+kMBZMREkp5CeXQYADlcv1NPbHX/94ygcTszFrp+y8ENC\njuFcfkkt8m/sX9jL2x1jh/XC2PDeLRa+8OqhYF0dWY2xn5fIgb44kXoN/z5+CSkXS7BixjD09O5Y\nnWtX2v3rZVwprEZ0RB9ER/SxdzhWp1QqUVv7+9Pa5iQQAFQqFdasWYNVq1bB29sbERER8PHxwbZt\n2yCRSBAXF4fz58/jhRdewPvvvw8/P7827yX2vRfF3H8x9x0Qd/+7e99LKjQ4fjYfvXzcMXfqELjI\npU3TQtH9+96VmAiSU0i/Wg53hQzBfR3vl1sqkSBmXH/8nJyPonJNq/O9fNwR+6fbjdYQdrYgWEwb\nxlLHGPsZixnX//c6122JWDQ1FHcM72u3LRrSs8vww8kc9PJxxyP3hNklBlsbPXo0jh8/jpiYGCQn\nJyMs7Pd+63Q6pKWlYfv27dBqtVi+fDlWr16NKVOmGK5ZvHgxXn/9dbNJIMAN5cXafzH3HRB3/8XQ\n9y9+OI+GRj2mRw9ARXnTh2q+egEyqaTb970tXCyGup2SSg2ul2swcnDPFqth+o6JdKii4OKKOqPH\nSyvruAcaOZRb61w/++48ki+WYEnMUKh62HYT+iq1Fh8fSIdUKsHjMyPgrhDHsDRt2jTExcVh4cKF\nAIDY2FgcOHAAGo0G8+fPBwDMnj0bCoUCy5Ytg7d3y0Wy+J5CRGJ1vUKDX88VorevB8ZHdv8ZJNYk\njhGXnFrGjW0jwoMdoz7QFFMb13ODdHJEhjrX/t7YdjADZy+W4FJ+AqIj+yDtSplFG9AnpBdZvGl9\nc5vm35lx4b0Q0lfVRb1zfBKJBK+99lqLYzdvBbFy5UqsXLnSZPsvvvjCarERETmyff+5Ap1ewKw7\nQlo8IGjvtmL0u+67JBt1G+lXm/cPdKz6wFtNjw42cZwLuZDj6unljr8sGoUFUwajRtOIQ4m5yCuu\nhV4QkFdciw/3pSEhvchk+4T0Iny4L63DbQzHMq632YaIiKigpBbxadfQz78HbgvvZe9wnB6fCJJD\nEwQBGdll8FK6IsDPsVcR5EIu5KykEgnuHdsfP501Xuf68f407DhywWjbGk2D0eMdaePoG90TEZF9\n7fnPFQgCMHviQEg5Rb7TmAiSQ8svrkWVugHREb2doiaGG6STMzNV56oXgB7uLkbPVamNJ3UdaePI\nG90TEZF9Xb1WjdPnryOkrwojQ3vaO5xugYkgOTRnmRZK1B2YqnPt56/E68vHGm2z7tOELmvDeloi\nulln6o8LStUI8DPfplP3sGKbjtyju9v962UAwJw7BzrFwwFnYNUaQUEQ8Oqrr2LhwoV49NFHkZub\n2+L8nj17MHPmTDzyyCPYtWtXi3OlpaWYNGkSrly5Ys0QycE17x9obCP5sqRUIDvbxhERdV8dqXO1\nVRsiEpfO1h/r9ebbdPoeVmrTkXt0d1n5lTh3qRRhQd4Y5uCLBzoTqz4RPHLkCLRaLXbu3ImUlBTE\nxsZi69atAIDy8nJs3rwZe/fuhVKpxNKlSzF+/HgEBASgsbERr776Ktzc3KwZHjm4Rp0embkV6OPr\nAV8VfxaIrK0jda62akNE4nIwPtvo8c+/P49fUgqMnrtcUGVRG0uvt1UbU9eLuY569y/mnwY62rZi\nzsCqiWBSUhImTpwIAIiKikJqaqrhXG5uLsLDw+Hp2bTx4fDhw5GcnIyAgAC88cYbWLRoET788ENr\nhkcO7kphFeq1OoffNoKoO+lInaut2hCReBSUqI0er2/QIeNG2Uh7WdrGFvfoSBux1lFnZJch42o5\nIkN8ERbkbb4BtZtVE8GamhpDogcAcrkcer0eUqkUwcHByMrKQllZGdzd3REfH4+QkBDs3r0bfn5+\nmDBhAj744ANrhkcOLv3G/oHDBrA+kIiISExM1yz3wPr/Ml5//Opnici3oI2l19uqjanre/m4G339\n7kwQBHx7ozZw9p0D7RxN92PVRFCpVKK29vcf5OYkEABUKhXWrFmDVatWwdvbGxEREfDx8cG2bdsg\nkUgQFxeH8+fP44UXXsD7778PPz+/Nu/l7+/Z5vnurLv2PaugClIJcMfoflB6uJq8rrv2vz3Yd/ES\ne/+JqHubHh2MD/elGT0ulRqfGjjDwjaWXm+rNqaur1ZrUVSuRm8fx95Oqyudu1SKS/lVGB3mj5C+\nKnuH0+1YNREcPXo0jh8/jpiYGCQnJyMsLMxwTqfTIS0tDdu3b4dWq8Xy5cuxevVqTJkyxXDN4sWL\n8frrr5tNAgGguLjaKn1wdP7+nt2y73XaRpzPLsOAPp7Q1NZDU1tv9Lru2v/2YN/F2XdA3P1nAkwk\nDqNCe0ImlUAikUAQBKvUHztqXbSx6wN6eiAx4zre2H4Gzy8a5VSrLHd0ldUD8dmGJ6ODApgEWoNV\nE8Fp06YhLi4OCxcuBADExsbiwIED0Gg0mD9/PgBg9uzZUCgUWLZsGby9W8775dKw4nUhtxI6vYDw\nNqaFsiiYiIioezqfUwGdXkDMuCA8NHlwu9s11x+39wMzR62LNnb9wIBc7Dx6EW9sP4O/LBqFfv5K\ni2Kwh+YVUJs1r4AKwOT349Y2APDvny7BV+XG2vIuZtVEUCKR4LXXXmtxLCQkxPD/K1euxMqVK022\n/+KLL6wWGzm25m0juEQwERGR+KRcKgEARA0yPytMLO65LQguMgm+PHwBb/7rLJ5bMBID+jj2LAlT\nq79+dTgTmbkVRs+dyjC+TYa5VVPLklKbZo2IdMZMR3BDeXJIGVfLIZdJMTjQy96hEBERkQ0JgoBz\nWSXwUMgxuB//DrjZ5NH9IJdJ8fn35/HWjrNYvWAkBjrotMk6bSPyS4yvdFpb14ifzuZb9HpiXTXV\nmpgIksOpqtUi93oNwgf4wNVFZu9wiIiIyIbyS2pRWlWPseG9ILuxyCD9bmJUAORyKT45kI63d57F\nM/OjHGpbBXVdA44m5eHwqVwIgvFr+vh6YOWc4UbPvff/zqGoXNPquDPVRToLJoLkcM7n3Ng2gtNC\niYiIRCcl68a00ME97RyJ44qO6AO5TIqP9qXhnW+S8fTcEQgPtu92W9VqLX48nYujSXnQ1OvQw02O\nPwzxx+nM4lbXPnhHCAJ6Gk/sZk0caGKV1QFdHrPYMREkh/N7fSD3DyQiIhKblEulkEiA4QNZH9iW\n24b2glwmwft7UvHurnOY9od+OHepFAWlagT4tX91TktW9DR2/ZD+3jiUmIPjZ/OhbdBD5eGCGZOC\nMWlUINwV8httrLsyK3UME0FyOOnZ5fBQyDGgd9sF0CwKJiIi6l5qNA24lF+JQYFeULq72Dschzcq\n1B+r5o7A5l3n8N3JHMPxjqzOaa6NqeulEgn0ggAfTwXm3tUfd0YFQHFTaY+tVmYlyzERJIdyvUKD\nkso6jA7zN7kxKxEREXVPv10uhSBwtVBLDB/oBx9PBUoq61qd23nsIipqjO/F/ENCjtHjptqYul4i\nAR69dwgmRPaFi9x+NZ3cVsxyTATJoWTcmBYaPoD1gURERGJjqA8cxPpAS5RVGU/2Kmu0+PpYlkWv\nZWkbAcCmChOYAAAgAElEQVSkkYEW3YMcAxNBcijp2VwohoiISIx0ej1SL5fBT6VAoD9XiLREQE8P\n5BW33l7Bz8sNf5waarTNv368iNKq1k8RTbUxdX0AV/N0WkwEyWHoBQEZV8vh46lAH18Pe4dDRERE\nNpSVVwl1fSPGRfSGRMLyEEtMjw42utLmvLsGYVSov9E22ga9RW1MXc/VPJ0XE0FyGHnXa1CjacCE\nyD4cAIiIiEQm5VIpANYHdkRHVtq0tA1X8+x+mAiSw2ieFhrezmmhLAomIiLqPs5dKoWrXIqh/Vke\n0hHNK236+3uiuJ0rqlu6OidX8+xe7Le0D9Et0q82LxTD/QOJiIjE5HqFBgUltRgW7AvXm7YeIGqv\nsqRUIDvb3mE4FSaC5BAadXpcyK1AQM8e8PFU2DscIiIisqFzN1YLHcFpoUQ2w0SQ7C4hvQgvf3wS\n2gY9qmq1SEgvsndIREREZEPnbtQHMhEksh3WCJJdJaQXtViBqkbTYPiac9CJiIi6vzptI87nlKN/\nLyV8VW72DodINPhEkOzqQHy20eMH46/aNA4iIiKyj/TscjTqBIwYzKeBRLbEJ4Jkc4Ig4FJBFRLS\ni5BvZPNTACgsNX78ZmVJqfD39wTauTIWEREROZ5zl5rqA6MG9bRzJETiwkSQbEIQBORer0FCehES\nM66jtKoOQNPuD3qh9fV9/XrYOEIiIiKyNb0gIOVSKTw9XBDSV2XvcMiJcVsxyzERpC6VkF6Eg/HZ\nKChRI6CnByYM7wtNfSMSM67jWpkaAOCukGFCZB+MHdYb1WotPjmQ0ep1pkcPsHHkREREZGs5RdWo\nrNFifGQfSKUSe4dDJCpMBKnL3LrwS15xLb4+lgUAcJVLcdvQXhgb3hsjBvnCRf77HkEyqRQH46+i\nsLQWff16YHr0AC4UQ0REJALnsppWC40azGmhRLbGRJC6zMH4bKPHfVUKbFwxDm6uxn/cxg3rzcSP\niIhIhFIulUAmlSAi2NfeoRCJDlcNpS5TUKI2eryyRmsyCSQiIiJxqqypx5XCaoT284KHG/9OILI1\nJoLUZQJ6ehg9bq2FX3zHRALBwVZ5bSIiIrKuc5c5LZTInpgIUpeZHh1s4jgXfiEiIqKWWB9IXaks\nKRXIzrZ3GE6FiSB1mXHDemNYsA+AptV7+/kr8fjMCNb/ERERUQsNjXqkZpehl487+vgan1FERNbF\nCdnUpapqtXB1keKfz9wJuYyfMxAREVFrF/IqUK/VIWoEnwYS2Qv/UqcuU6NpQF5xLQYFeDEJJCIi\nIpNSskoAAFGD/ewcCZF48a916jIX8yoAAEOCvO0cCRERETkqQRBwLqsUbq4yhPFvBiK7YSJIXeZC\nblMiaKs3dRYFExEROZ9rZWpcr9AgIsSXM4iI7Ii/fdRlLuRWQCaVYGCAyt6hEBERkYNKubFa6IhB\nnBZKXYfbilmOiSB1CU19I65eq0FIgAquLjJ7h0NEREQO6tylpvrAEYO4UAyRPVl11VBBELB+/Xpk\nZmbC1dUVmzZtQlBQkOH8nj17sG3bNqhUKsyaNQvz5s1DY2MjXnrpJeTn56OhoQFPPPEEpkyZYs0w\nqQtcKqiEXhAQ1o9z/YmIiMg4dV0DLuZVIqSvCl49XO0dDpGoWTURPHLkCLRaLXbu3ImUlBTExsZi\n69atAIDy8nJs3rwZe/fuhVKpxNKlSzF+/HicPHkSPj4+ePPNN1FZWYlZs2YxEXQCtq4PJCIiIueT\neqUMOr2AKE4LJbI7qyaCSUlJmDhxIgAgKioKqamphnO5ubkIDw+Hp6cnAGD48OFITk7Gfffdh5iY\nGACAXq+HXM6tDp3BhZwKSCTA4EAve4dCREREDqq5PjBqMKeFEtmbVbOsmpoaQ6IHAHK5HHq9HlKp\nFMHBwcjKykJZWRnc3d0RHx+PkJAQuLu7G9o+/fTTePbZZ60ZInWBhkYdLhdWoX8vT3i42S5x9x0T\nCUglwKnfbHZPIiKi7iYhvQgH47NRUKJGQE8PTI8Oxrhhvbu0TUJ6EQ7EZyO/uBZSqQSFpbUY0MfT\n5PVEZH1W/atdqVSitrbW8HVzEggAKpUKa9aswapVq+Dt7Y2IiAj4+PgAAAoLC7Fy5Uo88sgjuP/+\n+60ZInWBywVVaNQJnBZKRNROrKEnR5GQXoQP96UZvs4rrjV8bSqxs7TNrdfr9QI+2p8OiURiNuEk\naq+ypFT4+3sCxdX2DsVpWDURHD16NI4fP46YmBgkJycjLCzMcE6n0yEtLQ3bt2+HVqvF8uXLsXr1\napSUlGD58uVYt24dbr/99nbfy99fvJ8q2bvvx1IKAAC3RfaxbSxSCQD799+e2HfxEnv/nR1r6MlR\nHIzPNnr804MZOHDC+LmicrVFbUxdfzD+KhNBIjuyaiI4bdo0xMXFYeHChQCA2NhYHDhwABqNBvPn\nzwcAzJ49GwqFAsuXL4e3tzc2bdqEqqoqbN26FVu2bIFEIsEnn3wCV9e2V5YqFmn27+/vafe+nz1/\nHQDQS6WwaSy+egEyqcTu/bcXR/i3txcx9x0Qd/+7SwLMGnpyFAUlxpO0Rp0elbVaE+cEi9qYur6w\ntNbocSKyDauOIhKJBK+99lqLYyEhIYb/X7lyJVauXNni/Msvv4yXX37ZmmFRF9Lp9cjKq0RAzx5Q\neXAZaCKi9mANPTmKgJ4eyCtunZD181fi9eVjjbZZ92mCRW1MXd/Xr0cHIiairsIN5alTcopqUN+g\nY30gEZEF2ltD/5e//KVVDf2SJUswe/Zs1tBTl7hrZKDR49OjB5hsMz062KI2ll5PRLbBeSXUKZk5\nN/YP7Gf7bSNYFExEzoo19LYj5v63p+86NNXb+3m5oaK6HkG9PTH/7lDcOaqfyTYz7vKESuWGfx+9\niNyiarNtLL2+q/DfXpzE3HdLMRGkTuFG8kRElmMNvW2IvZ7WXN8bdXp8H38F7go5Ni4fB4WrzHDO\nXNvwfl5Yt+QPLY611cbS6zuL//bi67vvmMimtSNEvK2YpUkwE0HqML0g4GJeBXp6ucFX5WbvcIiI\nnAZr6MkRJF8sQWWNFlPH9GuRBBKROLBGkDqsoLgWtXWNGMKngURERE7n2Jk8AMDk0cbrBImoezOb\nCObn5+O//uu/cM899+D69et49NFHkZeXZ4vYyMFlclooEYkcx0hyVgUltTifU4Gh/b25eieRSJlN\nBNetW4fly5ejR48e8Pf3x4wZM/DCCy/YIjZycIb6wP5MBIlInDhGkrP66Ww+AGDKaOsu2EJEjsts\nIlheXo477rgDgiBAIpHgoYceQk1NjS1iIwcmCAIu5FXAS+mKXt7udonBd0wkEBxsl3sTEQEcI8k5\n1Wt1iEsthJfSFSNDe9o7HCKyE7OJoJubG65duwaJpGl54dOnT5tdoYy6v+sVGlTWaDEkyNvws0FE\nJDYcI8kZJWQUQVOvw11RAZDLuFwEdQ9lSalAdra9w3AqZlcNffHFF/H4448jJycHDz74ICorK/GP\nf/zDFrGRA7uQw/pAIiKOkeRsBEHAsaQ8SCUS3BkVYO9wiMiOzCaCpaWl2LVrF7Kzs6HT6TBw4EB+\n2km/1wf2YyJIROLFMZKczeWCKuRcr8GYMH9u/UQkcmbnA7z11ltwcXFBaGgohg4dygGOADStGNrD\nTY4Af640RkTixTGSnM3xG4vEcMsIIjL7RDAoKAgvvvgioqKi4Ob2+ydHs2bNsmpg5LjKqupQUlmH\nkYN7Qsr6QCISMY6R5Eyq1VokZlxHb18PhA/wsXc4RGRnZhNBH5+mN4qUlJQWxznIidcFB9k/sCwp\nFf7+nkBxtV3jICLx4hhJzuQ/vxWiUafH5FGBXOiNiMwngrGxsWhoaMCVK1eg0+kQGhoKudxsM+rG\nmhPBIdw/kIhEjmMkOQu9IOCns/lwlUsxYXgfe4dD1OV8x0QCUglw6jd7h+I0zI5WqampeOqpp+Dt\n7Q29Xo+SkhJs2bIFUVFRtoiPHFBmbgUULjL07620dyhERHbFMZKcRdqVMhRX1OGOEX3Rw83F3uEQ\nkQMwmwhu3LgRf//73w2DWnJyMjZs2IBdu3ZZPThyPFW1WhSWqhER4guZlHsPEZG4cYwkZ3H8TNMi\nMVO4SAwR3WD2L3m1Wt3ik82RI0eivr7eqkGR47qY5xj1gUREjoBjJDmDkkoNUi6VIKSvCsF9VPYO\nh4gchNlE0MvLC0eOHDF8/eOPP8Lbm0mAWF3IrQQADGEiSETEMZKcws/JBRAEYPIoPg0kot+ZnRq6\nYcMGPP/883j55ZcBNC2V/eabb1o9MHJMF3IrIJdJENLX096hsCiYiOyOYyQ5uoZGPX5NKUAPNznG\nhveydzhE5EDMJoLBwcF4//334eHhAb1ej9LSUgwYMMAWsZGDUdc1Iud6NUIDveAil9k7HCIiu+MY\nSY4u6cJ1VKkbcO/YILi6cOym7ovbilnO7NTQL774Ao899hg8PDxQWVmJJ554Al9//bUtYiMHk5Vf\nCUEAwrhtBBERAI6R5Ph+urFIzKSRnBZKRC2ZTQS/+eYbbN++HQAQGBiIb7/9Fl999ZXVA6Oul5Be\nhHWfJmDFG8ex7tMEJKQXWdTeUTaSJyJyFBwjyZHlXa/BhbxKRIT4orevh73DISIHY3ZqaENDA1xd\nXQ1fu7hw7xlnlJBehA/3pRm+ziuuNXw9bljvdr3GhdwKSCUSDArwskqMRETOhmMkObLjZ29sGcFF\nYojICLOJ4NSpU7FkyRLcd999AIDDhw/j7rvvtnpg1LUOxmebOH61XYlgfYMOVwqrMKCPEu4Ksz82\nRESiwDGSHJW6rgEn0q7Bx1OBEYP97B0OETkgs3/RP//88/jhhx9w6tQpyOVyPProo5g6daotYqMu\nVFCiNnq8sLS2Xe0vF1RBpxccalooi4KJyN44RpKj+ulMHuq1Otw/rj9kUrOVQEQkQmYTQa1Wi+Dg\nYMTExGDfvn1ISkrCiBEj0KsXlyB2Jj3c5ahWN7Q67qV0hSAIkEgkbbZnfSARUWscI8nRJKQX4WB8\nNvKKmz7oVbpzujKJA7cVs5zZj4ief/55HDp0COfOncOWLVugVCqxZs0aW8RGXWT/iWyjSSAAlFXV\n4/09qajRGD/frDkRDO3HRJCIqBnHSHIkzesBNCeBAPDl4QsWLw5HROJgNhHMy8vD008/jR9++AFz\n587Fn//8Z1RWVtoiNuoChxJzsPuXy/BTueHhqWHo56+ETCpBP38lFt09GIP7eeF0ZjFe+SQB5y6V\nGH2NRp0el/IrEejfg58sEhHdhGMkOZK21gMgIrqV2amhOp0OZWVlOHr0KN577z0UFxejrq7OFrFR\nJx0/k4evj2XBW+mK5xeNRC8fD9z9h34trrl7TBB+uJEsvvvvc7hrZAAWTBkMN9fffzSuXquGtlHP\naaFERLfgGEmOpLPrARCRuJhNBJcvX46HHnoIU6ZMQVhYGO699148/fTTtoiNOuE/5wrx5eELUHm4\n4PlFo9DLx/j+QVKpBPffPgDDB/rh4/3p+Dm5AOnZZVgxY5hhGmjztNAhTASJiFrgGEmOpJePO66V\ntU4G+/r1sEM0ROToJIIgCJY00Ol0kMlk1oqnw4pFunKkv79nq74npBfho/1p8FDI8cIfR6NfL2W7\nXquhUY+9/7mC7xOappDcN24AAnp64KvDF1Cn1aGvnwdmTghp976D1uY7JhIyqQTFIi0KNvZvLxZi\n7jsg7v77+3vaO4Q2OeIYKdafFUBcvyt6QcDaTxJwrbR1Ivj4zAiHGbttRUz/9rcSa9/F/nchYPkY\nafGGcJYMcIIgYP369cjMzISrqys2bdqEoKAgw/k9e/Zg27ZtUKlUmDVrFubNm2e2DbUtKbMYH+9P\nh5urDM8tHNnuJBAAXORSzJs0CFGD/fDJgXR8d7JlTUFhqdriTeiJiMTE0ZJAcizNK3oWlKgR0NMD\n06ODu2w8PZyYi2ulagT38USjTo/CUjX6+vXA9OgBHLNJFLitmOWsujP4kSNHoNVqsXPnTqSkpCA2\nNhZbt24FAJSXl2Pz5s3Yu3cvlEolli5divHjxyMtLc1kG2rbuUul+GBvKlzkUjw7fySC+6g69Dqh\n/bzx2rKxeH7rCdTWNbY6395N6ImIiKhJ84qezfKKa7vsw9W84hp8+8slqDxc8MxDUVB5uIr2qRAR\ntZ9VdxhNSkrCxIkTAQBRUVFITU01nMvNzUV4eDg8PT0hkUgwfPhwJCcnt9mGTMvILsOW3b9BKpXg\n6XkjMLifV6dez81VDk29zug5Fp0TERFZxlorejbq9PhkfzoadQKW3DcUKg/XTr0eEYmH2SeCv/76\nK/7+97+jqqoKgiAYNh8/evSo2RevqamBp+fvc1Xlcjn0ej2kUimCg4ORlZWFsrIyuLu7Iz4+HiEh\nIW22od8ZppeUquGnUqCsuh4SAE/NHYGhA3y65B4BPT1a7EXUjEXnRERNOjNGkrhYa0XPvf+5gpzr\nNbhjRF+MCvXv1GsRkbiYTQQ3btyINWvWIDQ0FBKJxKIXVyqVqK39/Q3u5oROpVJhzZo1WLVqFby9\nvREREQEfHx94enqabNMWR19AoCv9cjavxfSS4oqmpcpn3zUIk8cFd9l9Ft07FG99lWTk+BDH+H5L\nm34eHSIWO2HfxUvs/XcUnRkjSVxMf7hqfFXv9sjKr8R3J6+ip5cbFt0d2pnwiEiEzCaCPj4+mDx5\ncodefPTo0Th+/DhiYmKQnJyMsLAwwzmdToe0tDRs374dWq0Wy5cvx+rVq9HY2GiyTVvENA9+x6Hz\nRo+fSi/CA9EDuuw+4f288PjMCByMv4rC0lpD0Xl4Py/H+H6f+k3UNRDsuzj7Doi7/46WAHdmjCRx\nGRveG3nFl42eq2/QQeFi2UJD9VodPjmQDgjA8unhcFdYddkHIuqGzL5rjBkzBrGxsZg4cSIUCoXh\n+G233Wb2xadNm4a4uDgsXLgQABAbG4sDBw5Ao9Fg/vz5AIDZs2dDoVBg2bJl8Pb2NtqGmugFAVl5\nlcg38okiYJ3avXHDenNhGCIiEzozRpK4FFdoAAA9vdxQXl2Pvn4ekEokyLleg3e/ScFT80ZYlMx9\nczwL18s1iBnbH0P6d01JCJEz8x0T2TRbTMTbR1jK7DvOuXPnAADp6emGYxKJBF988YXZF5dIJHjt\ntddaHAsJCTH8/8qVK7Fy5UqzbcRMEARcLapGYvp1JJ4vQllVvclrWbtHRGRbnRkjSTxq6xqQkF6E\nnl5u+N8noiG9MY24UafHh/vSkJRZjHe+Scaz80fCw818Mvjb5VIcP5uPQP8emH1niNnriYiMMftu\n8+WXXwJoWvhFr9dDperYlgTUtlv3FoqO7Iu6+kYkZhShqLzpU0R3hRx3DO8LldIF38XntHqN6V04\nLZSIiMzjGEntEffbNWgb9Zg8OtCQBAKAXCbFEw9G4NODGTiZVoS3dp7FcwtGQunuYvK1ajQN2PZd\nBmRSCR6bMQwucu5dSUQdYzYRzM3NxbPPPovc3FwIgoCAgAC8++67CA4OtkF44mBsb6F/H88CALi6\nSDE2vBfGhfdG5EA/uMibFs4J8vdsVbvHKZxERLbFMZLM0QsCjp/Jg1wmxR3D+7Y6L5NKsWL6MLjI\npPj1XCHe/NcZ/GXhKKh6GN8G4qvDmais0WLuXQPRv7dj1cwSkXMxmwiuW7cOK1asQExMDADgu+++\nwyuvvGL4FJQ6p7y6HjuPXTR6zk+lwMYVt0Ph2vrTvubaPTEvGkFEZG8cI8mcjKvlKCrXYHxkH3ia\n2ONPKpVgyX1D4SKX4tiZfLxxIxn08VS0uO5k+jUkZlzH4EAv3DeOs4CIqHPM7stQXl5uGOAA4P77\n70dFRYVVg+ruajQN+Dk5v+lTvy1xqKzRGr2uokZrNAmkJr5jIgF+6k5EdsQxksw5lpQHAJg8OrDN\n66QSCR6eFoaYsf1RWKrGG9vPoKRSYzhfXl2Prw5dgMJFhhUzwiGVcrsSIuocs08EXV1dkZaWhoiI\nCABAamoq3N3drR6YM7u13m96dDCiBvvh7MUSJKQXIe1KGXR6AQAwuJ8XSio0qDCSDHLxFyIix8Yx\nktpSVlWH5KwSDOjtiYF9zdePSiQSzJ88CC5yKfafyMYb28/gntuC8Ou5QsMehBNH9EUvn47vPUjU\nXZUlpTZtMcSZcu1mNhF86aWXDJu+C4KAyspKvPPOO7aIzSkZq/f7cF8aZFKJIfnr30uJccN647bw\nXujp5d6qTTMu/kJE5Ng4RlJbfkougCAAU0YHQiJp3xM8iUSC2XcOhItcim9/uYwdR7NanP/1XCGG\nBftyXQAi6jSzieDIkSNx6NAhZGdnQ6/XIyQkBK6uxue4E3AwPtvocYkEmDkhGGPDeyOgZ8snfc1v\n5lz8hYjIuXCMJFMadXr8klIAD4UcYzswns8YH4yjZ/KMlo8cjL/KvxGIqNNMJoLvvfceVq1ahRdf\nfNHoeW70blx+ifFN3QUBmDVxoMl23LidiMh5cIwkc85cKEZVrRb33BYEhUvH6v2raxuMHi8sNf63\nBhGRJUwmgs31DmPHjm11rr3TG8SkoVGH//fzZQiC8fOs9yMi6j44RpI5x87kAwAmjWp7kZi2BPT0\nMNQG3ox/UxBRVzCZCE6ZMgUAcP36dTz++OMtzrH+oaWcomp8vD8d+SW18Orhgkojn+Cx3q/rsSiY\niOyFYyS1Ja+4BhdyKxAR7IM+vh1f2GV6dDDXECAiqzGZCL799tsoLS3FsWPHkJ2dbTiu0+mQkpKC\n1atX2yI+h6bXC/g+4Sr2/HoFOr2AyaMD8dCkwUjOKmG9HxFRN8Yxktpy/GzT08DJo/t16nW4hgBR\n+/mOiQSkEuDUb/YOxWmYTATvueceXLp0CSdPnmwx9UUmk+HJJ5+0SXCO7Hq5Gp8cyEBWfiW8lK5Y\ndn84hg/0A8B6PyKi7o5jJJmiqW/EidRr8PFUIGqwX6dfj39TEJG1mEwER4wYgREjRmDq1Knw9PQ0\nHBcEAXl5eTYJzhEJgoBfUgqw82gW6ht0+MPQXnj03iFQurvYOzQiIrKRzo6RgiBg/fr1yMzMhKur\nKzZt2oSgoCDD+T179mDbtm1QqVSYNWsW5s2bZ7YNOYb4tGuo1+pw/7j+kEml9g6HiMgks9tH7N27\nF++88w40Go3hWGBgII4cOWLVwBzFzZvD9/Z1h4tMipzrNXBXyPHYA8Nw+7DeXBiAiEikOjpGHjly\nBFqtFjt37kRKSgpiY2OxdetWAEB5eTk2b96MvXv3QqlUYunSpRg/fjzS0tJMtiHHIAgCjp/Jh0wq\nwZ1RAfYOh4ioTWYTwW3btmHv3r1499138eyzzyIxMRFxcXG2iM3ubt3ovbBUDQAI6NkDqx+Kgq/K\nzV6hERGRA+joGJmUlISJEycCAKKiopCammo4l5ubi/DwcMOTxuHDhyM5ORnnzp0z2YYcw4XcCuSX\n1GJseC94KRX2DoeIqE1m5yz4+fkhKCgIQ4YMwYULFzBnzhxcuXLFFrHZnanN4aUSMAl0AL5jIoHg\nYHuHQUQi1tExsqampsWUUrlcDr1eDwAIDg5GVlYWysrKoNFoEB8fD41G02YbcgzNi8RM6eQiMURE\ntmD2iaC7uztOnjyJIUOG4MiRIxg+fDiqqqpsEZvdFZSojR5vfjJIRETi1tExUqlUorb29/3h9Ho9\npDfqyVQqFdasWYNVq1bB29sbERER8PHxgaenp8k2bfH39zR7jSP75Wwe/n30InKKqtG/tyfm3x2K\nO0e1P9GyVf/Lq+qQlFmMAX08MX5UP4coG3H2f/vOEnP/Rdn3nKsAAH87h+FMzCaCa9euxa5du7Bm\nzRrs2rULMTExWLVqlS1isztu5EpERG3p6Bg5evRoHD9+HDExMUhOTkZYWJjhnE6nQ1paGrZv3w6t\nVovly5dj9erVaGxsNNmmLcVOvNfqrSUa2YVVeOurJFRV1bVrJU1/f0+b9X9/XNNWUneO6IuSkhqb\n3LMttuy7IxJz/9l3cfYdsPwDALOJYFhYGF566SUAwHvvvdexqJwUN3IlIqK2dHSMnDZtGuLi4rBw\n4UIAQGxsLA4cOACNRoP58+cDAGbPng2FQoFly5bB29vbaJvuzlSJxsH4qw61pYJOr8dPyQVQuMpw\ne0Qfe4dDRNQuJhPBKVOmtDmt4ejRo1YJyJGMG9YblbX12Hk0CxIAgf5KbuRKRESdHiMlEglee+21\nFsdCQkIM/79y5UqsXLnSbJvuTKfXI7+k9awcACgsNX7cXpIvlqK8uh6TRwfCXWH2M3YiIodg8t3q\nyy+/hCAI2LJlC4KCgjBnzhzIZDLs379fVPsIBvorAQAPTAjGrIkD7RwNERE5Ao6R1tOo0+NE6jV8\nF38VgmD8mr5+HrYNyozjZ5v+zSePCrRzJERE7WcyEQwMbHozy8zMbDH9ZNmyZZgzZ471I3MQpZV1\nAICeXu52joRuVZaU2jQXWsRzwYnIPjhGdj1tgw6/pBTgh8QclFXVQy6TIDzYBxnZ5a2u9fF0gyAI\nDrEgS2FpLdKzyxEW5I1+Nz48JiJyBu2av3Dy5EncfvvtAICff/4ZMpnMqkE5kpLKpk2C/b25XQQR\nEbUm5jHSUgnpRTgYn42CEjUCenpgenQwRgzyw09n83HoVC6qarVwdZHintuCcO/Y/vDxVNxocxWF\npbXo5eMOTX0jfrtciq8OX8DD94RBaudk8KezBQCAKaP5NJDInnzHRDbt8XbqN3uH4jTMJoIbN27E\nCy+8gOLiYgiCgMDAQLz55pu2iM0hlFQ0PRH082IiSERELYl9jLTErSuA5hXX4sN9aVC4SFHfoIe7\nQobp0QMw7bYgqDxcDdeNG9a7RW1+Va0Wb+9MxvGz+WjU6bEkZiikUtsngwnpRThwIhv5JbWQSiRo\n1HFPRyJyLmYTwWHDhmH//v0oLy+HRCKBt7e3LeJyGCWVdZBKJPDxVNg7FCIicjBiHyMtYWoF0IZG\nPR0F7l8AACAASURBVGZPDMHdY/rBw83F7Ouoerjir38chb99nYxfzxWiUafHsunhkLVjT8WucmtS\nqxcEfHIgAzKplAvKEZHTMJkIvvLKK9iwYQMWL15sdA7+F198YdXAHEVJpQa+KoVNBxgiInJsHCMt\nV1CiNn5CIsEDE0KMnzNB6e6C5xeOxN//nYL4tCI06AT86YFhkMtsM1bv/vWy0eOOtq0FEVFbTCaC\nCxYsAADRbB5vTEOjHhU1Wgztz094iYjodxwjLdfb1x2Fpa2TwQC/Hh16PQ83F6x+aCT+sescTp+/\njsZGPf57ViRc5NZLBi8VVOLgiau4Xq4xet7RtrUgImqLyXdLjUaDU6dOQSKRGP1PDMqquGKoI/Md\nEwkEB9s7DCISIY6RlqmoqYemvtHouenRAzr8uu4KOZ59KArDgn2QnFWC9749B22DrsOvZ0pmTjn+\ntvMsNn2RhOSsEri6GP/zqW8Hk1oiInsw+URw8+bNJhtJJBJRTHspvrFiaE8uFENERDfhGNl+ZVV1\neHPHWVTUaDF8oB/Kq+tQWKpGX78emB49oNNTKRUuMjw9bwS27E7FuUul+Meuc3hq7ggoXDu3eqsg\nCEjLLsOBuGxcyKsEAIQP8MED44NRUVOPj/ant2rTmaSWiDqH24pZrs0N5cWupJIrhhIRUWscI9un\nuEKDt3acRUllHaZHD8CcOwda5Ympi1yGP88ejg/2puLsxRK89nkipFIprpWpEeDXtE1FWwnnzdta\n9O3pgcgQX1zIrcCVwqY/KEcM8sOM8cEYHOhlaCORSAzbWnRVUktEZEtmVw09ffo0Pv30U6jVagiC\nAL1ej4KCAhw7dswW8dnV75vJMxEkIqLWxDxGmnOtTI23dpxFeXU9Zk8MsXhBGEu5yKX471mReGP7\nGVwqqDIcb96moqhMjYgQ31bt0q6UYc9/rhi+zi+uRX5xU63fmCH+mBEdjAF9PFu1u3VbCyIiZ2M2\nEVy7di0ee+wx7N69G4sXL8Yvv/yCYcOG2SI2uyupZI0gERGZJuYxsi35JbV4e8dZVNZq8dDkwYgZ\n198m95XLpKgzUSO45z9XWiR85vT2dcefZw/vqtCIiByO2UTQzc0Nc+fORX5+PlQqFTZu3Ig5c+a0\n68UFQcD69euRmZkJV1dXbNq0CUFBQYbz+/btw+effw6ZTIY5c+Zg0aJFaGxsxAsvvID8/HzI5XJs\n2LABISHW/RTRlJJKDWRS7iFIRETGdWaM7K5yiqrx9s5k1Gga8PC0MNw9pp9N719oYpsKiQRGE9If\nTuZAMHJ9SUVdF0dGRORYzCaCCoUCFRUVCAkJQUpKCqKjo6FWm9gL6BZHjhyBVqvFzp07kZKSgtjY\nWGzdutVw/s0338T3338PNzc3TJ8+HTNmzEBiYiL0ej127tyJEydO4O9//3ubRfnWVFJZB1+VAlIp\nV4BzRCwKJiJ768wY2R1dKazCO18nQ13XiCUxQ3DXyECbxxDQ0wN5xa23cQjsqcT8SYNbHf/tUqnR\n67kCKBF1d2Y321m6dCmeffZZTJ48GXv27MH06dMRGRnZrhdPSkrCxIkTAQBRUVFITU1tcX7o0KGo\nrKxEfX09gKbC6+DgYOh0OgiCgOrqari4uFjapy6hbdChskbLaaFERGRSZ8bI7uZiXgXe3nkW6vpG\nLJ8RbpckEACmRwebOG58RU9Lrycix8RtxSxn9onguHHjEBMTA4lEgm//f3t3Hh11leZ//FOVohJC\nhSUQbJtmVTAux3gAbWmHRpBMh4YGQlgCIjJNi1sfGSSALLIzgdj2yIDMwdP22CLHcOYAI9ADagho\nCwwgvw4tW1pb9tCQhQSykaXu74+YImkDSSBVlarv+/VXavvWc+vLqYenvvfeZ/NmnT59WtHR0Q06\neGFhoSIibiywdjgccrvdstur6s+ePXsqISFB4eHhio2NlcvlUmFhoc6fP6+4uDjl5+dr3bp1tzm0\nO5N7lR1DAQC3dic5MhhU77Z5IadIMpJs0vPDH9Rj9/tvE5XqDVwauqNnY58PAMGi3kIwPj5e0dHR\nGj58uAYPHtyoRfAul0tFRTemW9QsAjMzM7Vnzx6lp6crPDxcSUlJ2rlzpzIyMtS/f39Nnz5dly5d\n0qRJk7Rt2zY5nc5bvldU1Pd39LoT53Kregh2/WGbJj92U2vu8XmblcfP2K3L6uNvLu4kRwa6A8cv\nad3WY7XvNJKpa8Gdj1Xv6BkVFaHsBiwfYAdQAFZUbyG4e/du/d///Z+2b9+uN998U4899phGjBih\nfv361Xvw3r17a/fu3YqLi1NGRoZ69erleSwiIkItW7aU0+mUzWZTZGSkrl27pjZt2sjhcHieU1FR\nIbfbXe97NeSLvjH+djZPktTSYWvyYzelhia5YGXl8TN2a45dsvb4m1sBfCc5MtD9cf/pm9x/hqIK\nAAJAvYWg3W7XT37yE/3kJz/RgQMHtHLlSv3617/W4cOH6z14bGys9u7dq8TERElScnKytm/frpKS\nEo0ZM0Zjx47VhAkT5HQ61aVLF8XHx6usrExz587V008/rYqKCs2YMUNhYb6fnknrCABAfe4kRwa6\nrJvsznkx9/sbrwAAmp96C8Fjx45p+/btSktLU7du3fQv//Ivio2NbdDBbTabFi9eXOu+mq0gEhMT\nPUWiJyCHQ2+99VaDju9NOQVVU0NpJt98RfZ5SLLbpENf+TsUABZ1Jzky0N1sd0522wSAwFBvIfj6\n669rxIgR+vDDD9WhQwdfxNQs5BSUKsRuU1sXPQQBAHWzao6Uqnbb/N4aQbHbJgD/oK1Y49VbCG7e\nvNnz98aNGzVu3DivBtRc5BSUqn3rMHoIAgBuyqo5UqraYCX9/53X1+cLZLfb9EN22wSAgFJvH8Ga\nUlNTvRVHs1JWXqmrRWW0jgAANJhVcmRNRpLdZtO6pAFaMuUxikAACCCNKgRNc9gT2geqewiyPhAA\n0FBWyZE15eSXKLJ1qELsjfrvBACgGWjUN/c/buwSrLLzvysE27JjKACgYaySI6uVV1Qqv7CMH00B\nIEDVWwjm5+dr3759nr9feeUVffPNN14PzJ9y2TE0IOQdPiqdPu3vMABYmBVzZDVPmyV+NAWAgFRv\nIThjxgx9++232rdvnz755BMNGjRICxcu9EVsfnOjhyCFIADg5qyYI6tVz56JIlcCaAYi+zwkdevm\n7zACSr2FYEFBgSZOnKhdu3Zp5MiRGjlypEpKSnwRm9/QTB4A0BBWzJHVPP12uSIIAAGp3kLQ7Xbr\n6NGjSktL08CBA3XixAlVVlb6Ija/ySkolSPEpjYup79DAQA0Y1bMkdVyqq8IUggCQECqt4/gzJkz\nlZKSol/+8pfq3Lmzxo4dqzlz5vgiNr/JLSip6iFoo4cgAODmrJgjq2V/d0WQqaEAEJjqLQT79eun\nPn36yOl06syZM3rppZf02GOP+SI2v7heVqmrxeXq3NHl71AAAM2c1XJkTTn5pXI67GrditkzABCI\n6p0a+vbbb2v+/PnKysrS008/rT/84Q9asGCBL2Lzi5zvegi2Z31gs8eiYAD+ZrUcWVNOQYnatwmT\njdkzABCQ6i0Ed+3apWXLlmn79u0aPny4/uu//kvHjx/3RWx+QesIAEBDWS1HVisuLVdRaQXrAwE0\nG7QVa7wGbRbjdDq1e/duDRgwQG63O6h3RKN1BACgoayWI6tVt44gVwJA4Kq3EOzXr5+GDRum8vJy\nPfroo5o4caIGDRrki9j8gtYRAICGslqOrOZpHUGuBICAVe9mMbNnz9YzzzyjH/zgB7Lb7Xr99dd1\n//33+yI2v8jJr+6LxK+cAIBbs1qOrOZpJk+uBICAVe8Vwby8PK1cuVL9+vVT3759tWbNGuXk5Pgi\nNr+o6iHILmgAgPpZLUdWq74iyBpBAAhc9RaCCxYs0MMPP6xdu3YpPT1djzzyiObNm+eL2Pwip6BU\n7dvQQzAQsCgYgL9ZLUdWYxkFAAS+egvBc+fOacqUKXK5XGrdurWee+45ZWVl+SI2nystq1BhSTmL\n3wEADWKlHFlTdn6JWoU5FB5W7woTAPAJ2oo1Xr2FoM1m08WLFz23s7Ky5HAE5xd/LjuGAgAawUo5\nspoxRjkFpVwNBIAAV2+2mjZtmsaNG6eYmBgZY3TkyBEtXbrUF7H5XDaFIACgEayUI6sVFJWpvMLN\npmoAEODqLQTvvvtu/c///I/+8pe/yO12a/HixWrfvr0vYvO5XNY8AAAawUo5slpO9Y6h5EoACGj1\nFoLTp0/Xjh079OSTT/ogHP+60ReJXzkBAPWzUo6sll1AmyUACAb1FoL33nuv1qxZo5iYGIWF3fjS\nf/TRR70amD/kMDU0oET2eUiy26RDX/k7FAAWZaUcWc3Tb5crggAQ0OotBPPz83XgwAEdOHDAc5/N\nZtP777/v1cD8IaegVC0c9BAEADSMlXJkter19DSTB9Cc5B0+qqioCCn7mr9DCRj1FoLr169Xbm6u\n2rdvr5KSEl2+fFldu3b1RWw+l1tQqvatw2SjhyAAoAGslCOr3bgiSCEIAIGs3vYR69ev169+9StJ\nUl5enl544QVt3LjR64H5Wsn173oI8gsnAKCBrJIja8opKFVbl1MtHCH+DgUAcAfqLQQ3btyoDRs2\nSJI6deqkzZs364MPPvB6YL7GjqEAgMaySo6sVlHpVu7VUnVoS64EgEBX79TQ8vJyOZ031sy1aNHC\nqwH5CxvFAAAa63ZzpDFGixYtUmZmppxOp5YvX67OnTt7Ht+6davee+89hYSEaNSoURo/frwqKio0\ne/ZsXbhwQQ6HQ0uXLlX37t2bfEy3knftuoyRosiVABDw6i0EBw8erGeffVZDhgyRJH3yySd66qmn\nvB6Yr9E6IvCwKBiAv91ujkxLS1NZWZlSU1N15MgRJScna+3atZ7HU1JStGPHDoWFhWno0KEaNmyY\nDh48KLfbrdTUVO3bt0///u//rv/4j//w2tjqwo6hABA86i0EZ86cqZ07d+rQoUNyOByaNGmSBg8e\n7IvYfKr6imB7CkEAQAPdbo48fPiw+vfvL0mKiYnR0aNHaz0eHR2tgoICz+ZlNptN3bp1U2VlpYwx\nunbtml9m6Hhmz7CeHkAzQ1uxxqu3EJSkuLg4xcXFeTsWv6pOblH8ygkAaITbyZGFhYWKiIjw3HY4\nHHK73bLbq5bu9+zZUwkJCQoPD1dsbKxcLpcKCwt1/vx5xcXFKT8/X+vWrWvScTRE9ndXBDuyRhAA\nAl6DCsHbdTtrICTpnXfeUXp6usrLyzVhwgQlJCR4M0xJVVNDnQ67IsKDcw0kAKD5cLlcKioq8tyu\nWQRmZmZqz549Sk9PV3h4uJKSkrRz505lZGSof//+mj59ui5duqRJkyZp27ZttdYo1iUqKuKWjzfG\ntdIKSVKvHh0U1S68yY7rTU05/kBj5bFL1h6/Jcdur5pBYcmx3yavFoK3swbixIkT+vOf/6zU1FQV\nFxfr97//vTdD9MgtKFX7NvQQBAB4X+/evbV7927FxcUpIyNDvXr18jwWERGhli1byul0ymazKTIy\nUteuXVObNm3kcDg8z6moqJDb7a73vbKbcB31hUvXFGK3yZRVNOlxvSUqKiIg4vQGK49dsvb4rTr2\nSLdRiN1mybFXa2wR7NVC8HbWQHzxxRfq1auXXnrpJRUVFWnWrFneDFGSVFxaoaLSCvX4YRuvvxcA\nALGxsdq7d68SExMlScnJydq+fbtKSko0ZswYjR07VhMmTJDT6VSXLl0UHx+vsrIyzZ07V08//bQq\nKio0Y8YMhYX5dq1edn6J2rcOk93Oj6YAEOi8WgjezhqIK1euKCsrS+vWrdO5c+f04osvaufOnd4M\nU7lXaR0RiFgUDCBQ2Ww2LV68uNZ9NVtBJCYmeorEag6HQ2+99ZZP4qvL9bJKXS0u1486uvwWAwCg\n6Xi1ELydNRBt27bVPffcI4fDoe7duys0NFR5eXmKjIy85XvdyXzgby8VSpK6/rBNQM4rDsSYmwRz\nwRm7hVl9/PC9G22W2CgGQPNDW7HG82oheDtrIPr06aP169dr8uTJunTpkkpLS9WuXbt63+tO5gP/\n7ewVSVLLFvaAm1ds1XngEnPBrXzurTx2ydrjpwD2n+zq3bVpHQEAQcGrheDtrIFwOBz68ssvNXr0\naBljtHDhQq9v4OLpi8TUUAAA6kQzeQAILl4tBG9nDYQkJSUleTOs76me7kIzeQAA6ubpt0sPQQAI\nCnZ/B9Ac5BaUytnCroiW9BAEAKAu1c3kOzA1FACCAoWgqn7l7NCmJT0EA0ze4aPS6dP+DgMALCE7\nv1ShLUL40RQAgoTlC8Hi0nIVX69gfSAAADdhjFFOQYk6tA3jR1MAzVJkn4ekbt38HUZAsXwhyEYx\nAADcWlFphUrLKhXFRjEAEDQoBD2FIMkNAIC6eNYH8qMpAAQNCkGuCAIAcEueXMmOoQAQNCgEaR0B\nAMAtVfcQjCJXAkDQoBDM54pgoGJRMAD4RjY9BAEg6FAIFpQq1BkiF9thAwBQJ3oIAmjuaCvWeJYu\nBI0xyr1aog5t2A4bAICbyckvkatlC4U5Hf4OBQDQRCxdCBZfr1DJ9Up1aM0vnAAA1MVtjHKvliqK\nq4EAEFQsXQjeWB/ImgcAAOqSf+26KioNuRIAgoy1C8HvFr+zYygAAHW70TqCXAkAwcTiheB322GT\n3AISi4IBwPuyPa0juCIIAMHE4oUgU0MBALiVHFpHAAgAtBVrPEsXgrlMDQUA4JZoHQEAwcnShWBO\nQYnCnCFqFcZ22AAA1CUnv0Q2Se3ZYRsAgoplC0FjjHIKSukhCADALWQXlKpd61A5Qiz7XwYACEqW\n/VYvKq1QaVkl6wMBALiJ8gq38q9dJ1cCQBCybCFYvWNoB9YHBiwWBQOAd+VdLZWRFEWuBICgY91C\n0NNMnuQGAEBdsqt/NGXHUADNHG3FGs+6haBnx1CSGwAAdeFHUwAIXpYtBHMLSG4AANyKp5k8VwQB\nIOhYthC8Md2FQhAAgLpk00weAIKWZQvB3IJStQx1qFVYC3+HAgBAs5STXyJHiF1tXE5/hwIAaGKW\nLARr9hBE4GJRMAB4V05Bqdq3CZOdfrsAEHQsWQgWlpTrenklhSAAADdRcr1ChSXltI4AEBBoK9Z4\nliwEb+wYSnIDAKAu1bmS1hEAEJwsWQje2DGU5AYAQF1yqncM5UdTAAhKliwEPTuGktwAAKgTrSMA\nILhZshDMoYcgAAC3lO2ZGkquBIBgZMlCkGbywYFFwQDgPdVTQ1lGAQDByauFoDFGCxcuVGJioiZN\nmqRz587Venzr1q0aNWqUxowZow8//LDWY7m5uXryySd16tSpJo8rp6BU4aEOhdNDEACAOuUUlKpl\naIhahTn8HQoA1Iu2Yo3n1W/3tLQ0lZWVKTU1VUeOHFFycrLWrl3reTwlJUU7duxQWFiYhg4dqmHD\nhikiIkIVFRVauHChwsKa/opdVQ/BEv2gXXiTHxsAgGBgjFF2QYnuahcuGz0EASAoefWK4OHDh9W/\nf39JUkxMjI4ePVrr8ejoaBUUFOj69euS5Ek2K1eu1Pjx49WxY8cmj+mzjCyVlbt19nKhFrx7QAeO\nX2ry9wAAIJBdKy5XWbmbJRQAEMS8WggWFhYqIiLCc9vhcMjtdntu9+zZUwkJCfrFL36hJ598Ui6X\nS5s3b1b79u31xBNPyBjTpPEcOH5J73+c6bl9PrtI67YeoxgEAKCG6t212TEUAIKXV6eGulwuFRUV\neW673W7Z7VW1Z2Zmpvbs2aP09HSFh4crKSlJO3fu1ObNm2Wz2bR3716dPHlSs2fP1n/+53+qffv2\nt3yvqKiIWz4uSR8f+vIm95/TsAH3NmJkzUtDxh6U7FVXkC07fjF2K7P6+OFd2fm0WQKAYOfVQrB3\n797avXu34uLilJGRoV69enkei4iIUMuWLeV0OmWz2RQZGalr167pgw8+8DznmWee0ZIlS+otAiUp\nO/tavc85+/e6n3Pu0rUGvb45ioqKCNjY79ihryw9fsZuzbFL1h4/BbBv5ORX7a7NFUEACF5eLQRj\nY2O1d+9eJSYmSpKSk5O1fft2lZSUaMyYMRo7dqwmTJggp9OpLl26KD4+vtbrm3qB+g87hOt8dtH3\n7r+7fasmfR8AAAJZzndTQztQCAIIEJF9HqqaLXboK3+HEjC8WgjabDYtXry41n3du3f3/J2YmOgp\nEuvy/vvvN2k8Q/t107qtx+q4v2uTvg8AAIEsO59+uwAQ7CzVHOjHD9wlSfrj/jO6mFuku9u30tB+\nXT33AwCAqiuCrVs5FdoixN+hAAC8xFKFoFRVDFL4AQBQN7fbKO/qdXX7AesxASCYebV9BAAACCx5\n10pV6TasDwSAIEchCAAAPHJYHwgAlmC5qaEAAPibMUaLFi1SZmamnE6nli9frs6dO3se37p1q957\n7z2FhIRo1KhRGj9+vCTpnXfeUXp6usrLyzVhwgQlJCQ0eWzVPQRpHQEgkOQdPlrVYsii7ZVuB4Ug\nAAA+lpaWprKyMqWmpurIkSNKTk7W2rVrPY+npKRox44dCgsL09ChQzVs2DCdOHFCf/7zn5Wamqri\n4mL9/ve/90ps2QXf9RDkiiAABDUKQQAAfOzw4cPq37+/JCkmJkZHjx6t9Xh0dLQKCgo8/XRtNpu+\n+OIL9erVSy+99JKKioo0a9Ysr8RGD0EAsAYKQQAAfKywsFARETd25XQ4HHK73bLbq5bu9+zZUwkJ\nCQoPD1dsbKxcLpeuXLmirKwsrVu3TufOndOLL76onTt3NnlsOfmlsttsimwd2uTHBgA0HxSCAAD4\nmMvlUlFRked2zSIwMzNTe/bsUXp6usLDw5WUlKSdO3eqbdu2uueee+RwONS9e3eFhoYqLy9PkZGR\nt3yvqKjGtYHIu1aqDu1a6gd3tWn8wJqhxo4/mFh57JK1x8/Y0RAUggAA+Fjv3r21e/duxcXFKSMj\nQ7169fI8FhERoZYtW8rpdMpmsykyMlLXrl1Tnz59tH79ek2ePFmXLl1SaWmp2rVrV+97ZTdi44Sy\n8krlXb2u6C5tG/W65ioqKiIoxnE7rDx2ydrjZ+zWHLvU+CKYQhAAAB+LjY3V3r17lZiYKElKTk7W\n9u3bVVJSojFjxmjs2LGaMGGCnE6nunTpovj4eDkcDn355ZcaPXq0jDFauHChZw1hU8m9+l3rCNYH\nAggwkX0ekuw26dBX/g4lYFAIAgDgYzabTYsXL651X/fu3T1/JyYmeorEmpKSkrwal6d1BDuGAkDQ\no6E8AACQJGV/10yeHoIAEPwoBAEAgA4cv6SPvjglSdryp1M6cPySnyMCAHgTU0MBALC4A8cvad3W\nY57b2fklnts/fuAuf4UFAPAirggCAGBxf9x/+ib3n/FpHAAA36EQBADA4rJyiuu8/2JuUZ33A0Bz\nk3f4qHT6tL/DCCgUggAAWNwPO4TXef/d7Vv5OBIAgK9QCAIAYHFD+3W7yf1dfRsIAMBn2CwGAACL\nq94Q5o/7z+hibpHubt9KQ/t1ZaMYAAhiFIIAAEA/fuAuCj8AsBCmhgIAAACAxVAIAgAAAIDFUAgC\nAAAAgMVQCAIAAACAxVAIAgAAAIDFUAgCAAAAgMVQCAIAAACAxVAIAgAAAIDFUAgCAAAAgMVQCAIA\nAACAxVAIAgAAAIDFOLx5cGOMFi1apMzMTDmdTi1fvlydO3f2PL5161a99957CgkJ0ahRozR+/HhV\nVFRo7ty5unDhgsrLy/XCCy9o0KBB3gwTAAAAACzFq4VgWlqaysrKlJqaqiNHjig5OVlr1671PJ6S\nkqIdO3YoLCxMQ4cO1bBhw/Tpp5+qXbt2SklJUUFBgUaOHEkhCAAAAABNyKuF4OHDh9W/f39JUkxM\njI4ePVrr8ejoaBUUFMhms0mSbDabhgwZori4OEmS2+2Ww+HVEAEAAADAcrxaZRUWFioiIuLGmzkc\ncrvdsturlib27NlTCQkJCg8PV2xsrFwuV63XTps2TdOnT/dmiAAAAABgOV4tBF0ul4qKijy3axaB\nmZmZ2rNnj9LT0xUeHq6kpCR9/PHH+tnPfqaLFy/q17/+tSZOnKif//znDXqvqKiI+p8UpKw8dsna\n42fs1mX18aPhrP5vxcrjt/LYJWuPn7GjIby6a2jv3r312WefSZIyMjLUq1cvz2MRERFq2bKlnE6n\nbDabIiMjdfXqVeXm5mrKlCmaOXOm4uPjvRkeAAAAAFiSzRhjvHXwmruGSlJycrKOHTumkpISjRkz\nRqmpqdq0aZOcTqe6dOmipUuXauXKldqxY4d69OghY4xsNpt+97vfyel0eitMAAAAALAUrxaCAAAA\nAIDmh4byAAAAAGAxFIIAAAAAYDEUggAAAABgMQHbrb3mRjROp1PLly9X586d/R2WT40aNcrTe/FH\nP/qR/u3f/s3PEXnfkSNH9Jvf/Ebr16/X2bNn9dprr8lut6tnz55auHChv8PzuprjP3HihJ5//nl1\n69ZNkjR+/HgNGTLEvwF6QUVFhebOnasLFy6ovLxcL7zwgu69917LnPu6xn/33Xdb4ty73W7Nnz9f\np06dkt1u1+LFi+V0Oi1z7u+E1XOkFfOjZO0cacX8KFk7R1o5P0pNkyMDthBMS0tTWVmZUlNTdeTI\nESUnJ2vt2rX+DstnysrKJEnvv/++nyPxnd/97nf66KOP1KpVK0lVu9C++uqr6tu3rxYuXKi0tDQN\nHjzYz1F6zz+O/+jRo/rlL3+pyZMn+zcwL9u6davatWunlJQUXb16VSNGjFB0dLRlzn3N8RcUFGjk\nyJF6+eWXLXHu09PTZbPZ9OGHH+rgwYP67W9/K2OMZc79nbByjrRifpSsnSOtmh8la+dIK+dHqWly\nZMBODT18+LD69+8vSYqJidHRo0f9HJFvnTx5UsXFxZoyZYomT56sI0eO+Dskr+vatavefvttI8Hb\n3QAADO1JREFUz+1jx46pb9++kqSf/vSn2r9/v79C84m6xr9nzx5NnDhR8+bNU3FxsR+j854hQ4Zo\n2rRpkqTKykqFhITo+PHjljn3NcfvdrvlcDh07Ngx7d69O+jP/eDBg7V06VJJUlZWltq0aWOpc38n\nrJwjrZgfJWvnSKvmR8naOdLK+VFqmhwZsIVgYWGhIiIiPLcdDofcbrcfI/KtsLAwTZkyRe+++64W\nLVqkpKSkoB9/bGysQkJCPLdrdj5p1aqVrl275o+wfOYfxx8TE6NZs2bpgw8+UOfOnbV69Wo/Ruc9\nLVu2VHh4uAoLCzVt2jRNnz7dUuf+H8f/r//6r3r44Yc1e/bsoD/3kmS32/Xaa69p2bJlGjZsmKXO\n/Z2wco60Yn6UrJ0jrZofJWvnSKvnR+nOc2TAFoIul0tFRUWe2263W3Z7wA6n0bp166bhw4d7/m7b\ntq2ys7P9HJVv1TzfRUVFat26tR+j8b3BgwfrgQcekFSVBE+ePOnniLzn4sWLevbZZxUfH6+hQ4da\n7tz/4/itdO4lacWKFfr44481f/58Xb9+3XO/Fc797bJyjiQ/VrHa92RNVvuOtHKOtHp+lO4sRwZs\nVujdu7c+++wzSVJGRoZ69erl54h8a9OmTVqxYoUk6dKlSyoqKlJUVJSfo/KtBx54QIcOHZIkff75\n5+rTp4+fI/KtKVOm6KuvvpIk7d+/Xw8++KCfI/KOnJwcTZkyRTNnzlR8fLwk6f7777fMua9r/FY5\n9x999JHeeecdSVJoaKjsdrseeughHTx4UFLwn/s7YeUcSX6sYuUcaZXvSMnaOdLK+VFqmhxpMzWv\nIQaQmjuiSVWLort37+7nqHynvLxcc+bMUVZWlux2u5KSkvTII4/4Oyyvu3DhgmbMmKHU1FSdPn1a\nr7/+usrLy3XPPfdo2bJlstls/g7Rq2qO//jx41q6dKlatGihqKgoLVmyxLNQPpgsX75cO3bsUI8e\nPWSMkc1m07x587Rs2TJLnPu6xj99+nSlpKQE/bkvKSnRnDlzlJOTo4qKCj3//PPq0aOH5s+fb4lz\nfyesnCOtmh8la+dIK+ZHydo50sr5UWqaHBmwhSAAAAAA4PYE7NRQAAAAAMDtoRAEAAAAAIuhEAQA\nAAAAi6EQBAAAAACLoRAEAAAAAIuhEAQAAAAAi6EQREC4cOGCoqOjtX///lr3Dxo0SFlZWXd8/KY6\nzq1cvHhRQ4YMUUJCgoqLiz33FxYWKiEhQfHx8Tpz5kyjjztp0qSmDLPJXLhwQYMGDbqt186ZM0cX\nL170yXsBQCAjP94c+ZH8iFujEETAcDgcmj9/fq0k0VQNUn3RaPXAgQN68MEHtWnTJoWHh3vuP3Hi\nhJxOp7Zs2aKuXbs2+rgHDx5syjCb1O1+rgcOHFBjWpxWN5IFACsiP9aN/Eh+xK05/B0A0FAdO3bU\nE088oRUrVmjJkiWS5PkyPHjwoFavXq3169dLqvrF7Mc//rEeffRRvfzyy+rcubP++te/6qGHHtJj\njz2mLVu26OrVq1qzZo169OghY4xWr16tkydPKjQ0VIsXL9Z9992n3NxcLViwQH//+99lt9v16quv\nql+/flqzZo0yMjL097//XU8//bTGjx/vifP06dN6/fXXVVBQoPDwcM2bN08tWrTQqlWrVFxcrEWL\nFmnRokWSpLy8PM2bN085OTl66aWXtGbNGqWkpOjgwYNyu92Kj4/Xs88+q8rKSi1atEhff/21cnNz\n1b17d61evVpvvPGGJGncuHHauHGjoqOjdfLkSUnSli1bdPDgQSUnJ2vQoEGKiYnRyZMntWHDBn3+\n+ed6//33ZYzRgw8+qAULFshut2vu3Ln65ptvJEnjx4/XmDFjap2Dbdu26d1331VISIh+9KMf6Y03\n3pDT6dQ777yjnTt3yu1265/+6Z+UlJRU63U3+xwLCgo0b948ffvttwoNDdXs2bP1l7/8RZcvX9bU\nqVO1YcMGnTlzRitWrFBpaanatWunJUuWqFOnTjp+/Ljmz58vSbrvvvua8p8aAAQU8iP5kfyI22KA\nAHD+/HkzaNAgU1hYaAYOHGj27dtnjDFm4MCB5sKFC+bAgQPmmWee8Tz/tddeM1u2bDHnz5830dHR\n5sSJE8YYY2JjY81vf/tbY4wxq1evNsnJyZ7jrFu3zhhjzJ49e8zIkSONMcZMnz7dpKenG2OMuXz5\nshk8eLApKioyq1evrvV+NY0ePdp8+umnxhhjMjIyzMCBA01ZWZnZvHmzee211773/Jqxf/jhh2bF\nihXGGGOuX79uJk6caL788ktz6NAhs2TJEmOMMW6320ycONF88sknxhhj7rvvPs+xoqOjPX/XfL+B\nAweaLVu2GGOM+frrr82ECRPM9evXjTHGvPnmm2bt2rXm4MGDZurUqcYYY65cuVJnrE899ZTJzc01\nxhjz1ltvmRMnTpjPP//cvPLKK8btdhu3221mzJhhtm7d6jlnt/ocFy9ebFJSUowxxmRmZppx48Z5\n4s3KyjJlZWVm+PDh5uLFi8YYY/70pz+ZyZMnG2OMGTZsmNm/f78xxpi3337b814AYCXkR/KjMeRH\n3B6uCCKgtGrVSkuXLtX8+fO1devWBr0mKipK0dHRkqS77rpLjz/+uCSpU6dOtaaNjB49WpI0YMAA\nzZo1S4WFhdq3b59OnTqlVatWSZIqKyt19uxZSVJMTMz33qu4uFhnz57V4MGDPc9p27atTp061aBY\n9+3bp8zMTM9aj5KSEv31r3/V+PHj1bZtW23YsEGnTp3S2bNnVVRU1KBjVnv44YclVU0rOXPmjMaN\nGydjjCoqKvTggw9qwoQJOn36tKZMmaIBAwZo5syZ3zvGoEGDNH78eD311FOKi4tTdHS0PvroI331\n1VcaNWqUjDG6fv26OnXqpN69e9caV12f46FDh/Tmm29Kknr16qXU1FTPa4wxOn36tM6ePasXX3zR\n8+t2cXGxrly5osuXL3vO5ahRo7Rp06ZGfR4AEEzIj+RH8iMai0IQAeeJJ57QE088oZUrV3rmvdts\ntlpz5svLyz1/t2jRotbrHY66/9mHhIR873lut1t/+MMf1Lp1a0nS5cuX1aFDB6WlpSk0NPR7x3C7\n3XXeV1lZ2aCxud1uzZw505Mor1y5olatWmnXrl1avXq1Jk+erISEBF25cqXO19f8DCoqKmo9FhYW\nJqkqyQwZMkTz5s2TVJVMKysr5XK5tG3bNu3fv1979uzRyJEj9b//+79yuVyeY8ydO1ejR4/Wnj17\nNHPmTL388styu92aNGmSJk+eLKlqcX9ISIjy8vJqjavm55idna327dt/71x8++236t69u+d2ZWWl\nunTpoi1btnjGl5OT8731Dv947gDAisiP5EfyIxqDzWIQMGp+ic+aNUtffPGFLl++LElq166dzp8/\nr7KyMuXn5+vw4cN1vu5Wtm3bJkn69NNP1aNHD4WFhenxxx/Xhg0bJEnffPONhg8frtLS0psew+Vy\nqXPnzkpLS5MkZWRkKCcnRz179mxQDI8//rg2btyoiooKFRUVacKECTpy5Ij279+vn//85xo5cqQi\nIyN16NAhT/KsTsiSFBkZqW+++UbGGO3atavO93jssceUlpamvLw8GWO0cOFCvffee0pPT9fMmTM1\nYMAAzZs3T61ataq1M1llZaV+9rOfqV27dpo6daqGDx+uEydOqF+/fvroo49UXFysiooKvfjii/r4\n44+/N66an+MvfvELlZaWqm/fvvrjH/8oSfrb3/6m5557TjabTQ6HQ5WVlerRo4cKCgr05ZdfSpL+\n+7//WzNmzFDbtm3VqVMnffbZZ7XOHQBYEfmR/Eh+xO3giiACRs1fuVwul5YuXapf/epXkqR7771X\nAwYM0LBhw9SpUyf17du3ztfdbOcsm82m06dPa+TIkXK5XFq5cqUkaf78+VqwYIGGDx8uSfrNb35T\na0ezurzxxhtasGCBVq1apdDQUL399ts3/ZX1HyUmJurMmTOKj49XZWWlRo8erUcffVRt2rTRjBkz\ntHPnTjmdTj3yyCM6f/68pKrpKCNGjNCmTZv06quvaurUqerYsaN69+7t+WW05rijo6P18ssv69ln\nn5UxRvfff7+mTp0qu92uTz75REOHDlVoaKj++Z//uVaCDgkJ0bRp0zR58mSFhYWpTZs2WrFihTp2\n7KiTJ09q7Nixcrvd+ulPf6qRI0fqwoULntfe7HN85ZVXNH/+fI0YMUIOh8OzuP/JJ5/Uc889p3ff\nfVerVq3SsmXLVFZWVuvcpKSkaM6cOVq1apUeeeSRBn2+ABCMyI/kR/IjbofNNPTnIAAAAABAUGBq\nKAAAAABYDIUgAAAAAFgMhSAAAAAAWAyFIAAAAABYDIUgAAAAAFgMhSAAAAAAWAyFIAAAAABYDIUg\nAAAAAFjM/wdbPHh5TcqpZAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9d44f10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, 2, figsize=(15, 5))\n",
    "\n",
    "ax[0].plot(range(1, len(rfecv_svc.grid_scores_) + 1), rfecv_svc.grid_scores_, '-o')\n",
    "ax[0].set(xlabel='Number of features selected',ylabel='cross-validation score',title='SVC')\n",
    "ax[0].vlines(rfecv_svc.n_features_, \n",
    "             np.min(rfecv_svc.grid_scores_), \n",
    "             rfecv_svc.grid_scores_[rfecv_svc.n_features_-1], \n",
    "             color='red', linestyle='--')\n",
    "\n",
    "ax[1].plot(range(1, len(rfecv_lr.grid_scores_) + 1), rfecv_lr.grid_scores_, '-o')\n",
    "ax[1].set(xlabel='Number of features selected',ylabel='cross-validation score',title='logistic regression')\n",
    "ax[1].vlines(rfecv_lr.n_features_, \n",
    "             np.min(rfecv_lr.grid_scores_), \n",
    "             rfecv_lr.grid_scores_[rfecv_lr.n_features_-1], \n",
    "             color='red', linestyle='--')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Literature\n",
    "-----------\n",
    "\n",
    "<a name=\"scikit-feature-selection\">[1]</a> *\"scikit-learn documentation - 1.13. Feature selection\"*, [link](http://scikit-learn.org/stable/modules/feature_selection.html#feature-selection)\n",
    "\n",
    "<a name=\"pyml-ch4\">[2]</a> *\"Python Machine Learning Code examples - Chapter 4 - Building Good Training\n",
    "Sets – Data Preprocessing\"*, [link](https://github.com/rasbt/python-machine-learning-book/blob/master/code/ch04/ch04.ipynb)\n",
    "\n",
    "<a name=\"intro-ml-py-ch5\">[3]</a> *\"Introduction to Machine Learning with Python - Chapter 5 - Representing data and engineering features\"* \n",
    "\n",
    "<a name=\"diving-into-data\">[4]</a> *\"Diving into data - Selecting good features\"*, [part 1](http://blog.datadive.net/selecting-good-features-part-i-univariate-selection/), [part 2](http://blog.datadive.net/selecting-good-features-part-ii-linear-models-and-regularization/), [part 3](http://blog.datadive.net/selecting-good-features-part-iii-random-forests/), [part 4](http://blog.datadive.net/selecting-good-features-part-iv-stability-selection-rfe-and-everything-side-by-side/) "
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}