{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "#Agenda\n",
      "\n",
      "- Define the problem and the approach\n",
      "- Data basics: loading data, looking at your data, basic commands\n",
      "- Handling missing values\n",
      "- Intro to scikit-learn\n",
      "- Grouping and aggregating data\n",
      "- Feature selection\n",
      "- <p style=\"color: red\">Fitting and evaluating a model</p>\n",
      "- Deploying your work"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "#In this workbook you will\n",
      "- Create your own classifier for predicting delinquent customers\n",
      "- Build basic reports to help interpret the effectiveness of your model\n",
      "- Convert the results of your model to a credit score"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pandas as pd\n",
      "import numpy as np\n",
      "import pylab as pl"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "train = pd.read_csv(\"./data/credit-data-trainingset.csv\")\n",
      "test = pd.read_csv(\"./data/credit-data-testset.csv\")\n",
      "test.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>serious_dlqin2yrs</th>\n",
        "      <th>revolving_utilization_of_unsecured_lines</th>\n",
        "      <th>age</th>\n",
        "      <th>number_of_time30-59_days_past_due_not_worse</th>\n",
        "      <th>debt_ratio</th>\n",
        "      <th>monthly_income</th>\n",
        "      <th>number_of_open_credit_lines_and_loans</th>\n",
        "      <th>number_of_times90_days_late</th>\n",
        "      <th>number_real_estate_loans_or_lines</th>\n",
        "      <th>number_of_time60-89_days_past_due_not_worse</th>\n",
        "      <th>number_of_dependents</th>\n",
        "      <th>monthly_income_imputed</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.233810</td>\n",
        "      <td> 30</td>\n",
        "      <td> 0</td>\n",
        "      <td>    0.036050</td>\n",
        "      <td>  3300</td>\n",
        "      <td>  5</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 6017</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td> 1</td>\n",
        "      <td> 0.964673</td>\n",
        "      <td> 40</td>\n",
        "      <td> 3</td>\n",
        "      <td>    0.382965</td>\n",
        "      <td> 13700</td>\n",
        "      <td>  9</td>\n",
        "      <td> 3</td>\n",
        "      <td> 1</td>\n",
        "      <td> 1</td>\n",
        "      <td> 2</td>\n",
        "      <td> 2850</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.061086</td>\n",
        "      <td> 78</td>\n",
        "      <td> 0</td>\n",
        "      <td> 2058.000000</td>\n",
        "      <td>  2500</td>\n",
        "      <td> 10</td>\n",
        "      <td> 0</td>\n",
        "      <td> 2</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 2500</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.075427</td>\n",
        "      <td> 32</td>\n",
        "      <td> 0</td>\n",
        "      <td>    0.085512</td>\n",
        "      <td>  7916</td>\n",
        "      <td>  6</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 4145</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.046560</td>\n",
        "      <td> 58</td>\n",
        "      <td> 0</td>\n",
        "      <td>    0.241622</td>\n",
        "      <td>  2416</td>\n",
        "      <td>  9</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 2850</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 2,
       "text": [
        "   serious_dlqin2yrs  revolving_utilization_of_unsecured_lines  age  \\\n",
        "0                  0                                  0.233810   30   \n",
        "1                  1                                  0.964673   40   \n",
        "2                  0                                  0.061086   78   \n",
        "3                  0                                  0.075427   32   \n",
        "4                  0                                  0.046560   58   \n",
        "\n",
        "   number_of_time30-59_days_past_due_not_worse   debt_ratio  monthly_income  \\\n",
        "0                                            0     0.036050            3300   \n",
        "1                                            3     0.382965           13700   \n",
        "2                                            0  2058.000000            2500   \n",
        "3                                            0     0.085512            7916   \n",
        "4                                            0     0.241622            2416   \n",
        "\n",
        "   number_of_open_credit_lines_and_loans  number_of_times90_days_late  \\\n",
        "0                                      5                            0   \n",
        "1                                      9                            3   \n",
        "2                                     10                            0   \n",
        "3                                      6                            0   \n",
        "4                                      9                            0   \n",
        "\n",
        "   number_real_estate_loans_or_lines  \\\n",
        "0                                  0   \n",
        "1                                  1   \n",
        "2                                  2   \n",
        "3                                  0   \n",
        "4                                  1   \n",
        "\n",
        "   number_of_time60-89_days_past_due_not_worse  number_of_dependents  \\\n",
        "0                                            0                     0   \n",
        "1                                            1                     2   \n",
        "2                                            0                     0   \n",
        "3                                            0                     0   \n",
        "4                                            0                     0   \n",
        "\n",
        "   monthly_income_imputed  \n",
        "0                    6017  \n",
        "1                    2850  \n",
        "2                    2500  \n",
        "3                    4145  \n",
        "4                    2850  "
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "##Import some classifiers"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.linear_model import LogisticRegression\n",
      "from sklearn.neighbors import KNeighborsClassifier\n",
      "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n",
      "from sklearn.svm import SVC"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "##Fitting your model"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features = ['revolving_utilization_of_unsecured_lines', 'debt_ratio',\n",
      "            'monthly_income', 'age', 'number_of_times90_days_late']\n",
      "\n",
      "clf = KNeighborsClassifier(n_neighbors=13, warn_on_equidistant=False)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "-c:4: DeprecationWarning: The warn_on_equidistant parameter is deprecated and will be removed in 0.16.\n"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf.fit(train[features], train.serious_dlqin2yrs)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 5,
       "text": [
        "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
        "           n_neighbors=13, p=2, weights='uniform')"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###Generating Predictions"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#classes (returns an array)\n",
      "clf.predict(test[features])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 6,
       "text": [
        "array([0, 0, 0, ..., 0, 0, 0])"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#probabilities (returns a numpy array)\n",
      "clf.predict_proba(test[features])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 7,
       "text": [
        "array([[ 1.        ,  0.        ],\n",
        "       [ 0.84615385,  0.15384615],\n",
        "       [ 0.92307692,  0.07692308],\n",
        "       ..., \n",
        "       [ 0.84615385,  0.15384615],\n",
        "       [ 0.92307692,  0.07692308],\n",
        "       [ 1.        ,  0.        ]])"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###Plot a histogram of the probabilities"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "probs = clf.predict_proba(test[features])\n",
      "prob_true = probs[::,1]\n",
      "pl.hist(prob_true)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 8,
       "text": [
        "(array([  3.08820000e+04,   4.57100000e+03,   1.47800000e+03,\n",
        "         4.19000000e+02,   1.37000000e+02,   7.50000000e+01,\n",
        "         1.00000000e+01,   2.00000000e+00,   5.00000000e+00,\n",
        "         6.00000000e+00]),\n",
        " array([ 0.        ,  0.09230769,  0.18461538,  0.27692308,  0.36923077,\n",
        "        0.46153846,  0.55384615,  0.64615385,  0.73846154,  0.83076923,\n",
        "        0.92307692]),\n",
        " <a list of 10 Patch objects>)"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEACAYAAAC3adEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x107ef7cd0>"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "##[Evaluating your model](http://scikit-learn.org/stable/modules/model_evaluation.html#classification-metrics)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc, classification_report, confusion_matrix"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "preds = clf.predict_proba(test[features])\n",
      "preds"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 10,
       "text": [
        "array([[ 1.        ,  0.        ],\n",
        "       [ 0.84615385,  0.15384615],\n",
        "       [ 0.92307692,  0.07692308],\n",
        "       ..., \n",
        "       [ 0.84615385,  0.15384615],\n",
        "       [ 0.92307692,  0.07692308],\n",
        "       [ 1.        ,  0.        ]])"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "##Classification reports"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###What does this tell you?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "confusion_matrix(test['serious_dlqin2yrs'], clf.predict(test[features]))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 11,
       "text": [
        "array([[35051,    18],\n",
        "       [ 2494,    22]])"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print classification_report(test['serious_dlqin2yrs'], clf.predict(test[features]), labels=[0, 1])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "             precision    recall  f1-score   support\n",
        "\n",
        "          0       0.93      1.00      0.97     35069\n",
        "          1       0.55      0.01      0.02      2516\n",
        "\n",
        "avg / total       0.91      0.93      0.90     37585\n",
        "\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###Write your own version of confusion_matrix using `pandas`. Be sure to label the rows/columns.\n",
      "*HINT: use crosstab*"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pd.crosstab(test['serious_dlqin2yrs'], clf.predict(test[features]), rownames=[\"Actual\"], colnames=[\"Predicted\"])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th>Predicted</th>\n",
        "      <th>0</th>\n",
        "      <th>1</th>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>Actual</th>\n",
        "      <th></th>\n",
        "      <th></th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
        "      <td> 35051</td>\n",
        "      <td> 18</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td>  2494</td>\n",
        "      <td> 22</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 13,
       "text": [
        "Predicted      0   1\n",
        "Actual              \n",
        "0          35051  18\n",
        "1           2494  22"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###The ROC Curve\n",
      "To evaluate our classifier, we're going to use an [ROC Curve](http://en.wikipedia.org/wiki/Receiver_operating_characteristic). ROC Curves are great for evaluating binary (0, 1) classification models. A ROC Curve plots the False Postive Rate (fpr) vs. the True Positive Rate (tpr) for a classifier.\n",
      "\n",
      "See example in [scikit-learn docs](http://scikit-learn.org/stable/auto_examples/plot_roc.html#example-plot-roc-py)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def plot_roc(name, probs):\n",
      "    fpr, tpr, thresholds = roc_curve(test['serious_dlqin2yrs'], probs)\n",
      "    roc_auc = auc(fpr, tpr)\n",
      "    pl.clf()\n",
      "    pl.plot(fpr, tpr, label='ROC curve (area = %0.2f)' % roc_auc)\n",
      "    pl.plot([0, 1], [0, 1], 'k--')\n",
      "    pl.xlim([0.0, 1.05])\n",
      "    pl.ylim([0.0, 1.05])\n",
      "    pl.xlabel('False Positive Rate')\n",
      "    pl.ylabel('True Positive Rate')\n",
      "    pl.title(name)\n",
      "    pl.legend(loc=\"lower right\")\n",
      "    pl.show()\n",
      "plot_roc(\"Perfect Classifier\", test['serious_dlqin2yrs'])\n",
      "plot_roc(\"Guessing\", np.random.uniform(0, 1, len(test['serious_dlqin2yrs'])))\n",
      "\n",
      "#[::,1] selects the 2nd column of the numpy array\n",
      "plot_roc(\"KNN\", preds[::,1])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEZCAYAAACQK04eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10a154410>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEZCAYAAACQK04eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XlclFX7+PHPKGouaIgri6Hi44qAKaikYmpgLrmH5UIq\nj8ujaS5ZmUGWlaX2y0yT56nMfc99NzFzFwbcFwgVcQNBQFCH5fz+mJivCAgaw8BwvV+veb2YmTP3\nfd2IF4dzzn0djVJKIYQQosQoZeoAhBBCFC5J/EIIUcJI4hdCiBJGEr8QQpQwkviFEKKEkcQvhBAl\njCR+IQpBs2bN+OOPP0wdhhCAJH5hxlatWoW7uzuVKlWiZs2atG7dmoULF5okljNnztC+fXuTnFuI\nJ0niF2Zpzpw5TJgwgalTp3L79m1u377Njz/+yKFDh9DpdKYOTwiTksQvzE5CQgL+/v4sXLiQPn36\nULFiRQBcXFxYtmwZZcuWxdPTk59++snwmcWLF9OuXTvD8wsXLtClSxesra1p1KgRa9euNby3fft2\nmjZtSuXKlbGzs2POnDkAxMbG0r17d6ysrLC2ts7Sw3dwcOD3338HICAggAEDBjB06FAqV65Ms2bN\nCA4ONrQNCQnB1dWVypUrM2DAAN58802mT59unG+WKJEk8Quzc+TIER49esQbb7yRaxuNRoNGo8nx\nveTkZLp06cKgQYOIiYlh1apVjBkzhgsXLgAwfPhwAgMDSUxM5OzZs7z66quA/q8Me3t7YmNjuXPn\nDl9++WWW8z1uy5YtDBw4kISEBHr27MnYsWMB0Ol09O7dm2HDhhEfH8/AgQPZuHFjrrEK8Twk8Quz\nExsbS7Vq1ShV6v9+vNu2bYuVlRUVKlTg4MGDT/381q1bqVu3LkOHDqVUqVK4uLjQp08f1qxZA0DZ\nsmU5e/YsiYmJVKlSBVdXV8PrN2/e5MqVK5QuXRoPD49cz9GuXTu8vb3RaDQMGjSIsLAwAI4ePUp6\nejrjxo2jdOnS9O7dGzc3t3/6LREiC0n8wuxYW1sTGxtLRkaG4bXDhw8THx+PtbV1ltdzcvXqVY4d\nO4aVlZXhsWLFCm7fvg3A+vXr2b59Ow4ODnh6enL06FEApkyZgqOjI6+99hr169dn1qxZuZ6jZs2a\nhq8rVKjAw4cPycjI4MaNG9ja2mZpa29vj9RSFAVJEr8wO23atKFcuXJs3Lgx1zYVK1YkOTnZ8PzW\nrVuGr+vUqUOHDh2Ij483PJKSkvjhhx8AaNmyJRs3biQmJoZevXoxYMAAACpVqsTs2bOJiIhg8+bN\nzJ07l/379z9T7LVr1yY6OjrLa9euXZOhHlGgJPELs/Piiy/i7+/PmDFjWL9+PUlJSWRkZBAaGkpy\ncjIajQYXFxc2bNjAgwcPCA8PzzLR261bNy5dusSyZctITU0lNTWVEydOcOHCBVJTU1m+fDkJCQmU\nLl0aS0tLSpcuDeiHiMLDw1FKUblyZUqXLp1luCk/2rRpQ+nSpZk/fz5paWls2rSJEydOFOj3RwhJ\n/MIsTZkyhblz5/L1119Tq1YtatWqxahRo/j6669p06YN7733HmXLlqVmzZq88847DBo0yNCrtrS0\nZPfu3axatQpbW1tq167Nhx9+aFgGumzZMurWrUuVKlUIDAxk+fLlAISHh9OlSxcsLS1p27Yt//nP\nf+jQoUO22HKaWM58XrZsWTZs2MBPP/2ElZUVy5cvp3v37pQtW9aY3y5RwmhkIxYhijZ3d3fGjBnD\n0KFDTR2KMBPS4xeiiPnjjz+4desWaWlp/Prrr5w5cwZvb29ThyXMiIWpAxBCZHXx4kUGDBhAcnIy\n9evXZ926dVlWAQnxT8lQjxBClDAy1COEECVMsRjqcXFxMdzZKIQQIn86dOhAUFBQtteLRY8/LCwM\npVSJffj7+5s8Brl2uX65/uJ3/QcOHMgxpxaLxC+EEKLgSOIXQogSRhJ/MeDp6WnqEEymJF87yPXL\n9Xsa5bjFYjmnRqOhGIQphBBFSm6506g9/mHDhlGzZk2cnJxybfPuu+/SoEEDnJ2d0Wq1xgxHCCEE\nRk7877zzDjt37sz1/e3btxMeHs7ly5cJDAxk9OjRxgxHCCEERk787dq1w8rKKtf3N2/ebCg85e7u\nzr179wybXQghhDAOk07uRkdHY29vb3huZ2fH9evXTRiREEKYXno67N8Pd+4Y5/gmv3P3yYmH3HYa\nCggIMHzt6elZ4mf7hRDm6/vv4b334NAhqFEj/58LCgrK8U7dJ5k08dva2hIVFWV4fv369Wz7jWZ6\nPPELIYQ5UgqmToVvvoF166Bt22f7/JOd4k8//TTHdiYd6unZsydLliwB4OjRo7z44otSflYIUeLo\ndDpiY+/Svbs+6W/eDH37Gu98Ru3xDxw4kAMHDhAbG4u9vT2ffvopqampAIwcOZLXX3+d7du34+jo\nSMWKFfnll1+MGY4QQhQ5Wq2WIUN8SU7uR2TkdE6cgJYtjXtOuYFLCCFMQKfTMXPmTL77biEWFrOx\nshrM4cMaqlcvuHOY5AYuIYQQ2Wm1Wlq1asWqVcEkJITSt+8QLl0q2KT/NCZf1SOEECXN2rX7uXdv\nEteuDebYMQ1uboV7fkn8QghRSJYvh8BA+OOPifTuDSEhYG1d+HHIUI8QQhSChQth0CDo0gViYmDD\nBtMkfZAevxBCGI1Wq+XUqXvs3NmRVav0yzR79DB1VNLjF0KIAqfT6fjkE388PLzw9b1LcjKcOVM0\nkj5Ij18IIQqUVqtl4EBfIiPt0elCCQmxwdXV1FFlJT1+IYQoIPPnz6dDBy8uXpyEt/cWEhOLXtIH\n6fELIUSBSU1tQ1JSKN9+a8P48ZBLzUmTkzt3hRCiAGi10KIFfPWVvtBaUSB37gohhBGkpED//vqk\nP3Zs0Un6TyNDPUII8Qwya+yUKlWK117zp0cPSE6G4GB98i8OpMcvhBD5pNVqadmyFevXB7N9ux9t\n28LIkfpef3FJ+iA9fiGEyFNmL//bbxeSlDQbGMyYMRr27YNKlUwd3bOTxC+EEHkYMWIaO3eeJykp\nlB07bPD2NnVE/4ys6hFCiFw8egR+frB06QO8vF7g1181FKdNAnPLndLjF0KIJygF8+bBhAn650eP\nlsfd3bQxFSSZ3BVCiL/pdDrCw2/RurU+6a9fDxkZmFXSB0n8QggBwOHDWmrWbEWDBvNIT4c7d6BP\nn6J79+0/IYlfCFGi6XQ6+vTRV9KsUGESx4/P5ORJCm0bRFOQxC+EKLE2btRSo0YrfvstGD+/UKKj\nh9CqlRl28Z8gk7tCiBJp0SIYNSqYhg0nERIymHr1zD/hZ5LlnEKIEmfDBujbV78H7ltvmToa45Ei\nbUKIEu/8eejQQZ/0f/3VvJP+00jiF0KYvfnztdSvv5UmTfQF1a5fhyFDTB2V6UjiF0KYLZ1Oh5OT\nP+PGeWFnl0xcHJw8Cba2po7MtCTxCyHMklar5aWXWnHmTDCLF4dy4MCbWFmZOqqiQRK/EMLsBAYG\n0qGDF7duTWLr1i0MHWpj6pCKFFnOKYQwK0rBtm2vkJQUytatNnTrZuqIih5ZzimEMBsHD8KHH8Kh\nQxARAfXqmToi05LlnEIIs6SUIioKfH2hfXt45RV9nZ2SnvSfRoZ6hBDFkk6n47PPZhIamsTWrXNp\n3x4iI8HBwdSRFX2S+IUQxc7Jk1q6d/fl9m17IJB582DcOFNHVXwYdahn586dNGrUiAYNGjBr1qxs\n78fGxuLt7Y2LiwvNmjVj8eLFxgxHCFHM6XQ6Pv7Yn/btvbh9exL/+98WMjJsJOk/I6NN7qanp9Ow\nYUP27t2Lra0trVq1YuXKlTRu3NjQJiAggEePHvHll18SGxtLw4YNuX37NhYWWf8QkcldIQSAv38A\nCxacJD4+kCNHbGjVytQRFW2FvvXi8ePHcXR0xOHvATcfHx82bdqUJfHXrl2bU6dOAZCYmIi1tXW2\npC+EEErBunUwY8ZHQBkSEjRUrmzqqIovow31REdHY29vb3huZ2dHdHR0ljZ+fn6cPXsWGxsbnJ2d\n+e6774wVjhCiGPvqKxgwACZNKotSkvT/KaN1rzX52K/siy++wMXFhaCgICIiIujSpQthYWFYWloa\nKywhRDGg0+m4desWlSrVoV8/OHIEQkPB2dnUkZkHoyV+W1tboqKiDM+joqKws7PL0ubw4cNMmzYN\ngPr161O3bl0uXrxIy5Ytsx0vICDA8LWnpyeenp5GiVsIYVparZa33vKlXLlOhIXNpXZtuHIFatY0\ndWRFX1BQEEFBQXm2M9rkblpaGg0bNmTfvn3Y2Njg5uaWbXJ34sSJVKlSBX9/f27fvs3LL7/MqVOn\nqFq1atYgZXJXCLOn0+n4/POZfPvtQu7fn03z5oP54guNlFz4Bwp9ctfCwoL58+fj5eVFeno6w4cP\np3HjxixatAiAkSNH8tFHH/HOO+/g7OxMRkYGX3/9dbakL4Qwf1qtFl9fX5SyJzU1lA0bbOjd29RR\nmS+p1SOEMLnVq9cQFPSQH38czN69Gjp1MnVE5iG33CmJXwhhUgkJ0KOHfgL3hx/g3/82dUTmo9CH\neoQQIi9BQfr9b2vWhPh4qFTJ1BGVDFKdUwhRaLRaLatWrUKng7FjoVMnGDkSTp+WpF+YJPELIYxO\np9Ph7++Pl5cXR45kUK4c/PQT7NkDX3wBpUubOsKSRcb4hRBGlblip0YNfSXNvXtt+PprmDLF1JGZ\nP9mIRQhR6BYvXoyXlxdt2kxi794tXLpkw507kvRNTRK/EMJo2rVrj5dXKIsWDeG77zRcuQLVq5s6\nKiFDPUIIo3j0CFq2hDNn4OpVqFPH1BGVPDLUI4QwqscTTEqKfplmSook/aJIEr8Q4h/JXLHj5+cH\n6Auq2djA9ev6ZZqS9IseSfxCiOem1Wpp1aoVwcHBBATM4IsvoGlTePll0GqhQgVTRyhyku87d1NS\nUqgg/4pCCPS9/JkzZ7Jw4UJmz57Nm28OpmtXDUeOwNKl0K+fqSMUT5Nnj//w4cM0adKEhg0bAhAa\nGsqYMWOMHpgQouj6/vvvCQ4OJjg4lIyMIVSqpOHiRbh1S5J+cZDnqh43NzfWrVvHG2+8gVarBaBp\n06acPXu2UAIEWdUjRFGTlpZGXFxpXn1Vw9mz8O67IDunFj3/qEhbnSdmZ2RDdCFKNgsLC959FypX\n1lfXlD1wi5c8M3idOnU4dOgQoB/XmzdvXpZdtIQQ5kun03H16lUaNGhgeC0tDYYPh9WrISpKkn5x\nlOcY/8KFC/nhhx+Ijo7G1tYWrVbLDz/8UBixCSFMKHPFzrfffmt47d496NMH/vwTLl+GJ7bRFsVE\nnmP8hw4dwsPDI8/XjEnG+IUoPPq9bz/nxx9/ZM6cOQwaNAiNRoNOB/XrQ8WKcPKklFEuDp77zt2x\nY8fm6zUhRPGn1Wpp2bIlWq2W0NBQBg8ejEajISwMypXTD/OEhUnSL+5yHeM/cuQIhw8fJiYmhrlz\n5xp+ayQlJZGRkVFoAQohCs+tW7eYMmWKoZcPsGUL9OwJnp6wf79p4xMFI9fEr9PpSEpKIj09naSk\nJMPrlStXZt26dYUSnBCicHXt2jXL89BQfdKfOhW++spEQYkCl+cY/5UrV3BwcCikcHImY/xCFK4r\nV2DyZFi/Hvz9ISDA1BGJ5/Hc6/grVKjA5MmTOXfuHA8ePDAc7Pfffy/4KIUQhSIkJISQkBBGjBiR\n7b3YWGjRAurW1W+G3qFD4ccnjCvPyd23336bRo0a8ddffxEQEICDgwMtW7YsjNiEEAVMp9PxySef\n4O3tTfny5bO9v369fqMUW1s4dEiSvrnKM/HfvXuXESNGULZsWTp06MAvv/wivX0hiqGQkJAsK3be\nfvttw3u//w5t2ujr7CxYoC+n/MILJgxWGFWeib9s2bIA1KpVi61btxISEkJ8fLzRAxNCFJzly5fj\n7e3NlClT2Lx5MzY2Nob3vv0WvL3B3V0/tj96tOniFIUjz8ndLVu20K5dO6Kiohg3bhyJiYkEBATQ\ns2fPwopRJneF+Idu3LgBkCXhp6fr78LdvBm2b4cnFvQIM5Bb7nyuPXePHz+Om5tbgQSWH5L4hSh4\n8+fDpEnw11/6MX1hfp55VU9GRga//fYbERERNGvWjNdff52TJ0/y0UcfcefOHUJDQ40asBDi+WRk\nZFCqVO6juDqdfqnm99/DgQOS9EuiXHv8I0aMIDIyEjc3Nw4cOEDt2rW5cOECM2fO5I033jDc1Vco\nQUqPX4g8ZdbYuXTpEqtWrcqxTVIStGsH4eH6IZ5XXy3kIEWheuYe/9GjRzl16hSlSpXi4cOH1KpV\ni4iICKytrY0aqBDi2YWEhODr60udOnUIDAzMsc2WLTBsmH65ZkwM5LCaU5QQuf49WKZMGcOfiy+8\n8AJ169aVpC9EEaPT6fD398fb25vJkyezZcuWLBO4oJ/E/eQTfemF4cP1SzUl6Zdsufb4L1y4gJOT\nk+F5RESE4blGo+HUqVPGj04I8VQ///wzwcHBhIaGZkn4GRn60snLl8OKFfq7cXfuBC8vEwYrioxc\nx/ivXLny1A8WZv0eGeMXImcZGRloNJosc26rVsHQofpJXC8v6NVLP8Tz9y05ogQp0OWc+bVz504m\nTJhAeno6I0aMYOrUqdnaBAUF8d5775Gamkq1atUICgrKHqQkfiHyZfp0+PxzffJ/801TRyNMrdAT\nf3p6Og0bNmTv3r3Y2trSqlUrVq5cmWW/3nv37uHh4cGuXbuws7MjNjaWatWq5Tt4IUoKnU7H5cuX\nadq0aY7vK6Vfnjl+vL78QseOhRygKJKeeweu53X8+HEcHR1xcHCgTJky+Pj4sGnTpixtVqxYQd++\nfbH7e+POnJK+ECVdaGgobm5uzJ07N8f3U1P1vfvx4/U9fUn6Ii/5SvwpKSlcvHjxmQ4cHR2Nvb29\n4bmdnR3R0dFZ2ly+fJm4uDg6duxIy5YtWbp06TOdQwhzlrli57XXXmPixIn873//y9YmIwNGjtSX\nT751S4Z3RP7kWY9/8+bNTJkyhUePHnHlyhW0Wi3+/v5s3rz5qZ/Lzw1eqamphISEsG/fPlJSUmjT\npg2tW7emQYMG2doGPLYThKenJ56ennkeX4ji6tSpUwwZMgQ7O7tsK3YynT6tT/rHjsGpU1CzpgkC\nFUVKUFBQjvOkT8oz8QcEBHDs2DE6/v33o6urK3/99VeeB7a1tSUqKsrwPCoqyjCkk8ne3p5q1apR\nvnx5ypcvT/v27QkLC8sz8Qth7hISEpg4caJhs/MnzZmjL7swdKh+mWblyiYIUhQ5T3aKP/300xzb\n5TnUU6ZMGV588cWsH3pKHZBMLVu25PLly1y5cgWdTsfq1auzVfR84403+PPPP0lPTyclJYVjx47R\npEmTPI8thLlr164dQ4YMyZb0IyKgdWuYMgXWrIHFiyXpi2eXZ4+/adOmLF++nLS0NC5fvsy8efNo\n27Zt3ge2sGD+/Pl4eXmRnp7O8OHDady4MYsWLQJg5MiRNGrUCG9vb5o3b06pUqXw8/OTxC9EDsLD\n9WvxDx4EDw9ISZGNUsTzy3M5Z3JyMjNnzmT37t0AeHl5MX36dF4oxJ86Wc4pzFVoaChBQUFMmDAh\n1zZ370K1atC2LWzapP9aiPx47nX8ISEhtGjRwmiB5YckfmFudDodM2fOZOHChcyZM4fBgwdna3Pn\njn5t/uef6zdJ2b7dBIGKYu251/FPnDiRRo0aMX36dM6cOWOU4IQoSTLX5WfW2Mkp6W/YAE2b6pP9\nqlWS9EXByjPxBwUFsX//fqpVq8bIkSNxcnLis88+K4zYhDA769evN6zLz6mS5sWL+o1R+vaFCRMg\nOFjW5ouC90wlG06fPs2sWbNYvXo1qampxowrCxnqEebi7t27PHr0KMfSyXPmwMcf64d1VqyAihVN\nFKQwG889xn/u3DnWrFnDunXrsLa25s0336Rfv37UqFHDaME+SRK/MFfJyfD++7BokT75HzwIr7xi\n6qiEuXjuxN+6dWt8fHzo378/tibanFMSvyiO0tPTKV26dI7vJSfrNzv/4AOoWxc+/RR694ZKlQo5\nSGHWTFKWuaBI4hfFSeaKnZMnT7Jt27Ys78XFwdy5MHOmfgvE99+HSZOgELewFiXIM++5279/f9au\nXZtlF67HDyY7cAmRXWhoKL6+vtjZ2fHf//43y3srV8Jbb8FLL0FgIPj5mShIUeLl2uO/ceMGNjY2\nXL16NdtvDI1Gw0svvVQoAWaeT3r8oih7fF3+7NmzDTV2UlL0CX/WLLh8Gb77Dt5919TRipLimdfx\nZ646WLBgAQ4ODlkeCxYsMF6kQhRDa9euNazLz6yxs3WrfmXO2LH6LRATEyXpi6IhzzF+V1dXtFpt\nltecnJw4ffq0UQN7nPT4RVGX+fOp0WhIS4Ovv4Zp02DZMnj7bRMHJ0qsZx7jX7hwIQsWLCAiIiLL\nOH9SUhIeHh7GiVKIYkqj0fDgAfz6q34nrDJl9Hfbdu1q6siEyC7XHn9CQgLx8fF88MEHzJo1y/Bb\nw9LSEmtr68INUnr8oojQ6XScOXMmW/0qrRZ69QKdTr9iZ9gwEwUoxGOeeTlnYmIilStX5u7duzlu\nBFG1atWCjzIXkvhFUZC5YqdZs2YsW7bM8HpQkH4Mv39/WLIE8rFdhRCF4pkTf7du3di2bRsODg45\nJv7IyMiCjzIXkviFKeW2YicxEZyc4No1/Tj+0qWyHl8ULXIDlxDP4fTp0wwePBg7OzsCAwMNq90O\nHoT27aFsWYiOlhr5omh67rLMhw4d4v79+wAsXbqUiRMncvXq1YKPUIgiKD09PUslzbg4CAjQJ313\n9//bJEWI4iTPHr+TkxNhYWGcPn0aX19fhg8fztq1azlw4EBhxSg9flEkrF4N48ZBhQr6ompeXqaO\nSIine+4ev4WFBaVKlWLjxo385z//YezYsSQlJRklSCGKogsXoEED8PGBqVMhMlKSvije8kz8lpaW\nfPHFFyxbtozu3buTnp5eqLX4hSgMoaGh2TYYSk+HBQugcWNwcICHD6WgmjAPeSb+1atXU65cOX7+\n+Wdq1apFdHQ0U6ZMKYzYhDA6nU6Hv78/r732mqH+1LVrMGKEvtzCN9/otz7cswfKlTNxsEIUkHyt\n6rl16xYnTpxAo9Hg5uZWqJuwgIzxC+N4vJJm5oqdefP0d9526KDf5NzDQ3r4ovh67jH+NWvW4O7u\nztq1a1mzZg1ubm6sXbvWKEEKUVi2bduWZe/bKlVsaN9en/S3btXflPXKK5L0hXnKs8ffvHlz9u7d\na+jlx8TE0KlTp0Ktxy89flHQkpKSSEpKMizRrF4dLC3hxAn9RK4Q5uC5e/xKKapXr254bm1tLUlY\nFHuWlpa88IIN338PLVtC/foQHy9JX5QMuVbnzOTt7Y2XlxdvvfUWSilWr15NVyk5KIqR1NRUypQp\nk+W14cPh55/1N2GNHq1fny/DOqKkyNfk7oYNG/jzzz8BaNeuHb179zZ6YI+ToR7xPDJr7AQFBREU\nFIRGo2HXLv2KnRs39OP47dqZOkohjOeZ6/FfunSJKVOmEB4eTvPmzfnmm2+ws7MzapBCFJTHV+ys\nXLmS+HgN48frN0YZMQJmz4YqVUwdpRCmkesY/7Bhw+jevTvr16+nRYsWvCt7xoli4PF1+ZkrdoKC\nbLC21k/crl4N//2vJH1RsuXa479//z5+fn4ANGrUCFdX10ILSojntWvXLoKDgwkODiUiwoY+fWDj\nRv0duKNHmzo6IYqGXBP/w4cPCQkJAfQrex48eEBISAhKKTQaTbYdiIQoCrp37871692pU0c/U/va\na3DqlL5uvhBCL9fJXU9PzywbsGQm/Ez79+83fnR/k8ldkV/Hj+tX6nzxBXzwgazUESWbbMQizIpO\np+PkyZO0bdsW0FfQnDtXP37/7rvw3XcmDlCIIuC5b+D6J3bu3EmjRo1o0KABs2bNyrXdiRMnsLCw\nYMOGDcYMR5iJ0NBQ3Nzc+Pbbb8nIUHz2mb6C5sGDcP68JH0h8pLnDVzPKz09nbFjx7J3715sbW1p\n1aoVPXv2pHHjxtnaTZ06FW9vb+nVi6d6cu/bTp0G4+Oj4bff4PBhaNPG1BEKUTwYrcd//PhxHB0d\ncXBwoEyZMvj4+LBp06Zs7b7//nv69euXpSyEEE86d+4cbm5uBAcH8+23ocybNwQ7Ow0REXDnjiR9\nIZ5Fnok/IyODpUuXMmPGDACuXbvG8ePH8zxwdHQ09vb2hud2dnZER0dna7Np0yZG/73OTiMzcSIX\nZcuWZeLEiQwbtoVBg2zw8YGzZyE4GKysTB2dEMVLnol/zJgxHDlyhBUrVgBQqVIlxowZk+eB85PE\nJ0yYwFdffWWYgJChHpEbe3tHTp0aQt++GgIDYfJkaNLE1FEJUTzlOcZ/7NgxtFqt4QauqlWr5mvr\nRVtbW6KiogzPo6KispV8CA4OxsfHB4DY2Fh27NhBmTJl6NmzZ7bjBQQEGL729PTE09MzzxiEeVAK\n3ngD9u3T19fp0MHUEQlRNGXWpcqTyoObm5tKS0tTLi4uSiml7ty5Y/j6aVJTU1W9evVUZGSkevTo\nkXJ2dlbnzp3Ltb2vr69av359ju/lI0xhJrRarZo8ebLKyMhQSil1965Sbm5KgVLXr5s4OCGKmdxy\nZ55DPePGjaN3797cuXOHjz76CA8PDz788MM8f6FYWFgwf/58vLy8aNKkCW+++SaNGzdm0aJFLFq0\nKO/fSKJEebzGjtNjt9n27QsPHkBCAtjamjBAIcxIvm7gOn/+PPv27QOgU6dO2ZZkGpvcwGXectr7\nFuCzz/Sbnd+6BRUqmDhIIYqh575z99q1awCGD2dO2tapU6egY8yVJH7ztW/fPgYOHMjs2bMZPHjw\n3//W8Omn+seJE/odsoQQz+65E3+zZs0Myf7hw4dERkbSsGFDzp49a5xIcyCJ33w9evSIu3fvGnr5\nV6+CqyuULq2fzG3e3MQBClGMPfNGLJnOnDmT5XlISAg//PBDwUUmSrRy5cphY2PDgwcwdqx+O8Qu\nXWDXLil771IBAAAgAElEQVSwJoSxPPOduy1atODYsWPGiEWYuYcPH+bwGkydqh/DX7UKTp+G3bsl\n6QthTHn2+OfMmWP4OiMjg5CQEGxleYV4Bpk1drZt28aJEycMQ4fh4eDhAeXLy/p8IQpTnj3++/fv\nGx46nY7u3bvnWHNHiJxkVtIMDg5m8+bNhqR/4gQ0aABdu0JkpCR9IQrTU3v86enpJCYmZun1C5Ef\nT1bSzFyxs3s3BAbC+vUQEAD+/qaOVIiSJ9fEn5aWhoWFBYcOHcq2+5YQeTly5AghISGEhoZiY2ND\nTAy8/Tbs2QM9e0J0NPy9kEcIUchyXc7ZokULQkJCGDVqFDdu3KB///5U+PsuGo1GQ58+fQovSFnO\nWWylp+s3Of/vf6FmTQgJkYQvRGF55uWcmY0fPnyItbU1v//+e5b3CzPxi+JHKVi+HEaMAEtLuHYN\nHqvSLYQwoVwTf0xMDHPnzs1SN0WIJ+l0Og4ePEinTp0Mr6WmQv/+sHkzzJ4N48ZBmTImDFIIkUWu\niT89PZ2kpKTCjEUUM5k1durWrUvHjh0pVaoUWi0MG6Z/PzERKlUybYxCiOxyHeN3dXVFq9UWdjw5\nkjH+oiWnFTug4ZNP4PPPoV072LZNP8QjhDCd5y7ZIMTjLly4gI+PD3Z2doYVO0ePwvTpcPIkhIVJ\nfR0hirpcb+Dau3dvYcYhionKlSszadIktmzZwr17Nnh66jc6r18fbtyQpC9EcZCvevymJkM9RYtS\nsHYtvPkmeHvDnDmy/60QRVFuufOZi7SJki0jAwYN0if9yZNhxw5J+kIUN5L4RY5CQ0MZNWoUGRkZ\nhtdiYqBhQ1i9Gi5c0O+OJYQofiTxiywe3/u2bdu2hlIdmzZB7dr6Mfy4OP0vACFE8SSreoTB43vf\nZq7YUQoWLID//AcmTIBvvzV1lEKIf0oSvwDg8OHD9OrVK0slzfPnwd0ddDpYulQ/ti+EKP5kVY8A\n9Hdqx8TEUKtWLZSCAQNg3TqoUUO/SUrjxqaOUAjxrJ57s/WiQBJ/4VFKX3Jh8WIprCZEcSfLOYVB\ncnJyjq+HhYGzsz7ph4dL0hfCXEniL0EyV+y4ubmRnp6e5b0ffwQXF6hTB+7f19+JK4QwTzK5W0I8\nvmJnz549lC5dmtRUWLIEdu+GNWvg6FH9ZK4QwrxJj9/MPb4uf+LEiWzZsgUbGxu+/14/rDNmDLRs\nCWfPStIXoqSQHr+ZO336NKGhoYZ1+SEhMGOG/oasr7+Gd96BatVMHaUQojDJqp4S4t49/Wbn27fr\nd8d6/319T18IYb6kHn8JlZEBn36q7+XXqqWvtyM9fCFKNhnjNxM6nY6tW7dme334cH3S37hRXy9f\nkr4QQhK/GQgNDcXNzY3AwEDS0tIAePAAXF31a/KPHIE33oC/660JIUo4SfzF2JMrdjZt2oSFhQVJ\nSdCtGzx6BA8fQuvWpo5UCFGUyBh/MRUeHk6/fv2yVNIE/Zr8ESP0Cf/kSShXzsSBCiGKHKP3+Hfu\n3EmjRo1o0KABs2bNyvb+8uXLcXZ2pnnz5nh4eHDq1Cljh2QWrK2tef/99w3r8tPS9DV2vLygZ0+4\ndUt/F64QQjzJqMs509PTadiwIXv37sXW1pZWrVqxcuVKGj9W6vHIkSM0adKEKlWqsHPnTgICAjh6\n9GjWIGU5Z66Ugl27YPRofVG14GB96QUhhDBJkbbjx4/j6OiIg4MDZcqUwcfHh02bNmVp06ZNG6pU\nqQKAu7s7169fN2ZIZqd/f+jaFXr31k/oStIXQuTFqIk/Ojoa+8dKPNrZ2REdHZ1r+59++onXX3/d\nmCEVO6GhoQwaNIjU1NQsr1+6BK+8AuvXw927MHculC1roiCFEMWKUSd3Nc+wfnD//v38/PPPHDp0\nKMf3AwICDF97enri6en5D6Mr2nQ6HTNnzmThwoXMnj0bCwv9P1VkJIwcCXv2QOfOcP06VK1q4mCF\nEEVCUFAQQUFBebYzauK3tbUlKirK8DwqKgo7O7ts7U6dOoWfnx87d+7Eysoqx2M9nvjNXU573x4/\nDrNmwYYN0K+ffvK2Zk1TRyqEKEqe7BR/+umnObYz6lBPy5YtuXz5MleuXEGn07F69Wp69uyZpc21\na9fo06cPy5Ytw9HR0ZjhFAtarTZLJc07d2wYMADatIGKFfUTuWvXStIXQjw/o/b4LSwsmD9/Pl5e\nXqSnpzN8+HAaN27MokWLABg5ciQzZswgPj6e0aNHA1CmTBmOHz9uzLCKNBcXF86ePUtcXHVeegmi\novQTuLJaRwhRUKQ6ZxGTkQHffw8TJoCfH8yfL5O2QojnI9U5i6CEhATDUlaA1avBxwcsLWHHDvD2\nNmFwQgizJbV6TCCzxk6LFi3Q6XTcuAGffaZP+nPmQGKiJH0hhPFIj7+QabVafH19sbe35+OPDzJ4\ncFnWrAF7e32Pf8AAU0cohDB3kvgLSdZ1+XOIixvEsGEahg/X34zVoIGpIxRClBSS+AtJREQEZ86c\nYdOmUP79bxvOnIGDB/V33wohRGGSMf5CEBUFP//cmPDw9bRta0O9enD5siR9IYRpSI/fyM6cgRYt\n9DdcTZ2qL5sswzpCCFOSxF/AdDodGzduokKF/uzfD4sWwdChEBgoWx8KIYoGSfwFSKvVMmiQLzEx\ndYiJ6cmwYeVYuRJ69DB1ZEII8X8k8RcAnU7HRx/N5IcfFvLw4RwaNRrEpUsaXnzR1JGVLFWrViU+\nPt7UYQhR6KysrIiLi8t3e0n8/1BYWCSdOvXi7t062NmFsmWLjdTUMZH4+PgSU9pDiMc9Swl8kFU9\n/8i+feDiUoMqVT5i167NREZK0hdCFH2S+J/DwYP6O2w7d4aPP65IRMSbvPaaBgv5+0kIUQxI4n8G\ncXHQqRO0bw/370NIiL7GjhBCFCeS+PNp4UItNWr0ITb2IfHxsH07uLqaOiohhHh2kvjzEB6uo3Zt\nf8aM8cLHpzehoeVktY4QBeDcuXO0atXK1GEUC1u2bMHHx6fAjieJPxfx8dCrl5YGDVrx4EEIp06F\nsmzZ4GeePRcik4ODAxUqVMDS0pJatWoxePBgEhMTs7Q5fPgwr776KpUrV+bFF1+kZ8+enD9/Pkub\nxMREJkyYwEsvvYSlpSWOjo6899573L17tzAv5x+bPn06U6ZMMXUY/8iVK1fo2LEjFStWpHHjxuzb\nty/Pz+h0Oho3boy9vX2W1w8fPoybmxuVK1fG2dmZQ4cOGd7r0aMHZ8+e5fTp0wUStyT+HBw8CC1a\nXGTrVi8mT55MfPxmnJxsTB2WKOY0Gg1bt24lKSmJsLAwTp8+zeeff254/8iRI3h5edG7d29u3rxJ\nZGQkzs7OeHh4EBkZCeiTRqdOnTh//jy7du0iKSmJI0eOUK1aNaNuWZqWllagx7t58yZBQUH06tXr\nuT6fnp5eoPE8r4EDB/Lyyy8TFxfHzJkz6devH7GxsU/9zDfffEONGjWydCLj4uLo0aMHU6dOJSEh\ngffff58ePXpw7969LOcKDAwsmMBVMVBYYd6+rdRbbykFSvn7K3X3blyhnFcUjKL+4+zg4KD27dtn\neD5lyhT1+uuvG56/8sor6j//+U+2z3Xt2lUNGTJEKaXUf//7X1WzZk2VnJyc7/OeOXNGde7cWVWt\nWlXVrFlTffnll0oppYYOHao+/vhjQ7v9+/crOzs7w/OXXnpJzZo1Szk5Oaly5cqpWbNmqX79+mU5\n9rvvvqveffddpZRS9+7dU8OGDVO1a9dWtra26uOPP1bp6ek5xvTrr7+qLl26ZHntyy+/VPXr11eW\nlpaqSZMm6rfffjO898svv6i2bduq9957T1lbW6vp06erR48eqUmTJqk6deqomjVrqlGjRqkHDx4o\npZSKj49X3bp1U9WrV1dWVlaqe/fu6vr16/n+nuXHxYsXVbly5dT9+/cNr7Vv3179+OOPuX7mr7/+\nUo0bN1Y7duzI8r3esmWLatKkSZa2//rXv9RPP/1keH7o0CFVt27dHI+b289+bq9Lj/9vN26Ara2+\nqNqlSxAQAFWrWpk6LGFm1N83mF2/fp2dO3fi7u4OQEpKCkeOHKF///7ZPjNgwAD27NkDwN69e+na\ntSsVKlTI1/mSkpLo3Lkzr7/+Ojdv3iQ8PJxOnToB+r9A8hq6XLVqFTt27CAhIQEfHx+2b9/O/fv3\nAX2ve+3atbz99tsA+Pr6UrZsWSIiItBqtezevZv//e9/OR739OnTNGzYMMtrjo6O/PnnnyQmJuLv\n78+gQYO4ffu24f3jx49Tv3597ty5w0cffcTUqVMJDw8nLCyM8PBwoqOjmTFjBgAZGRkMHz6ca9eu\nce3aNcqXL8/YsWNzvc7u3btjZWWV46Nnz545fubs2bPUq1ePihUrGl5zdnbm7NmzuZ5n3LhxfPnl\nl7zwwgu5tsmUkZGR5ViNGjXiypUrhu//P5Lrr6YixJhhHjumlLd3rAKlunRRKiPDaKcSRpafnxMo\nmMfzeOmll1SlSpWUpaWl0mg0qlevXoYecVRUlNJoNOrixYvZPrdjxw5VpkwZpZRSnTt3Vh9++GG+\nz7lixQrVokWLHN/z9fV9ao/fwcFB/fLLL1k+88orr6glS5YopZTavXu3ql+/vlJKqVu3bqly5coZ\netyZ5+7YsWOO5/bz81MffPDBU2N3cXFRmzZtUkrpe/x16tQxvJeRkaEqVqyoIiIiDK8dPnw41x6x\nVqtVVlZWTz3fs1qyZIlq3bp1ltemTZumfH19c2y/YcMGw194T36vY2NjlZWVlVq1apXS6XRq8eLF\nqlSpUmrUqFGGNjqdTmk0GhUVFZXt2Ln97Of2eont8cfHw3vv6XB39ycoyJV9+1LYvVsqaJq7gkr9\nz0Oj0bBp0yYSExMJCgri999/5+TJk4C+1kqpUqW4efNmts/dvHmT6tWrA1CtWjVu3LiR73NGRUVR\nr1695wsYsk1AvvXWW6xcuRKAFStWGHr7V69eJTU1ldq1axt6yqNGjSImJibH41pZWZGUlJTltSVL\nluDq6mr4/JkzZ7JMWD8eS0xMDCkpKbz88suG9l27djWMr6ekpDBy5EgcHByoUqUKHTp0ICEhoUBL\nelSqVCnb5Py9e/eoXLlytrbJycm8//77fPfddzkey9ramo0bNzJnzhxq1arFrl276Ny5M3Z2doY2\nmd+vFwtgWWGJS/wJCTBtGlStqmXBgla0bx9CRMRRXn01f386C1EQ2rdvz7hx45g6dSoAFStWpE2b\nNqxZsyZb2zVr1hiGZzp37syuXbtISUnJ13nq1KnDX3/9leN7FStWzHKcW7duZWvz5FBQv379CAoK\nIjo6mo0bN/LWW28B+qRcrlw57t69S3x8PPHx8SQkJOS6CqV58+ZcunTJ8Pzq1av8+9//5ocffiAu\nLo74+HiaNWuWJVE/Hku1atUoX748586dM5zv3r17hkQ8Z84cLl26xPHjx0lISODAgQMopXJN/F27\ndsXS0jLHR7du3XL8TNOmTfnrr7+yDL2EhYXRtGnTbG0vX77M1atXadeuHbVr16Zv377cvHmT2rVr\nc+3aNUD/M3H8+HHu3r3LkiVLuHDhAm5uboZjnD9/HgcHBypVqpRjPM8k9z9kio6CCDM6Wqlx45SC\nR6pSpU9U5crV1ZIlS1SGjO2YjaL+4/zk5G5MTIyqUKGCOnr0qFJKqT///FNVrFhRzZs3TyUmJqq4\nuDg1bdo0ZWVlpcLDw5VSSj169Ei1atVKeXt7qwsXLqj09HQVGxurZs6cqbZv357tnElJSap27drq\n//2//6cePnyoEhMT1bFjx5RS+oniRo0aqbi4OHXz5k3l7u6ebajn8Xgzde3aVXXu3DnbENIbb7yh\nxo8frxITE1V6eroKDw9XBw4cyPF7cevWLWVtba0ePXqklFLq7Nmz6oUXXlAXL15UaWlp6ueff1YW\nFhaGyc1ffvlFvfLKK1mOMX78eDVgwAB1584dpZRS169fV7t27VJKKfX++++rrl27qocPH6q7d++q\nXr16KY1Gk+tk8/Nq3bq1mjx5snrw4IFav369evHFF1VsbGy2dmlpaer27duGx4YNG5SNjY26ffu2\nIaaQkBCl0+lUQkKCGj9+fLbrnTlzZo6T/0rJUE82qamwbRu0agXr1sHChTfp2vUC58+HMniwrMsX\nplOtWjWGDh3KrFmzAPDw8GDXrl1s2LABGxsbHBwcCAsL488//6R+/foAlC1blr1799KoUSO6dOlC\nlSpVcHd3Jy4ujtatW2c7R6VKldizZw9btmyhdu3a/Otf/yIoKAiAwYMH4+zsjIODA97e3vj4+OTr\n/8Nbb73Fvn37DL39TEuWLEGn09GkSROqVq1K//79c/wrAqBmzZq8+uqrbNy4EYAmTZowadIk2rRp\nQ61atThz5gyvPLY3aU4T0bNmzcLR0ZHWrVtTpUoVunTpYvgrYsKECTx48IBq1arRtm1bunbtapT/\n66tWreLkyZNUrVqVadOmsX79eqytrQE4ePAglpaWAJQuXZoaNWoYHlZWVobXSpXSp+FvvvmG6tWr\nU6dOHW7fvs1vv/2W7VwjR44skLg1f/9WKNI0Gs1zjc3FxECNGvqvx46FefNkDN+cPe/PiTCN8+fP\nM3ToUKPef2AutmzZwvLly1m1alWO7+f2s5/r6+aa+A8cAE9P8PGBv+eihJmTxC9KqmdN/GY51LN3\nL3h66ujU6VdWrJBEIIQQjzOrxH/3LoweDV26aKlVqxXly6/L9+oHIYQoKcxm65Dz56FLFx0azUyq\nVFnI11/PYdCgQTJ5K4QQTzCLxL9jB7z99nWU6oaHRx0CA0OxsZGiakIIkZNiPbmbng5jxkBgIEyc\nqKN1663069dbevkllEzuipLqWSd3i22P/+FDGDIE1q6FP/8ED4+yQB9ThyVMyMrKSn7pixLJyurZ\nCkoadXJ3586dNGrUiAYNGhhuUnnSu+++S4MGDXB2dkar1ebruFu3Qr16EBQEly+Dh0cBBi2Krbi4\nOMNt+fKQR0l6xMXFPdP/FaMl/vT0dMaOHcvOnTs5d+4cK1euzLaT0Pbt2wkPD+fy5csEBgYyevTo\nPI/7v/9p6dGjK926JXLtGjg6GusKio7MOy1LopJ87SDXL9cfZJTjGi3xHz9+HEdHRxwcHChTpgw+\nPj5s2rQpS5vNmzczdOhQANzd3bl3716W+tuP0+l0DB3qz8iRXnTv/haBgZbko6S1WSjJP/wl+dpB\nrl+uP8goxzXaGH90dHSWMqp2dnYcO3YszzbXr1+nZs2a2Y5nZ9eKmJg6DBwYyrJlNlJ6QQghnpPR\nevz5nWRTKuuMc26fc3CYxKJFm1mxwoZSZnXbmRBCFDJlJEeOHFFeXl6G51988YX66quvsrQZOXKk\nWrlypeF5w4YN1a1bt7Idq379+gqQhzzkIQ95PMPD2dk5x/xstKGeli1bcvnyZa5cuYKNjQ2rV682\n7NyTqWfPnsyfPx8fHx+OHj3Kiy++mOMwT3h4uLHCFEKIEsdoid/CwoL58+fj5eVFeno6w4cPp3Hj\nxixatAiAkSNH8vrrr7N9+3YcHR2pWLEiv/zyi7HCEUII8bdiceeuEEKIglOkpkmNdcNXcZHX9S9f\nvhxnZ2eaN2+Oh4cHp06dMkGUxpGff3uAEydOYGFhwYYNGwoxOuPLz/UHBQXh6upKs2bN8PT0LNwA\njSyv64+NjcXb2xsXFxeaNWvG4sWLCz9IIxk2bBg1a9bEyckp1zYFnvcKZCa3AKSlpan69euryMhI\npdPplLOzszp37lyWNtu2bVNdu3ZVSil19OhR5e7ubopQjSI/13/48GF17949pZRSO3bsMJvrz8+1\nZ7br2LGj6tatm1q3bp0JIjWO/Fx/fHy8atKkiYqKilJK6ffrNRf5uX5/f3/1wQcfKKX01161alWV\nmppqinAL3B9//KFCQkJUs2bNcnzfGHmvyPT4C/qGr+ImP9ffpk0bqlSpAuiv//r166YItcDl59oB\nvv/+e/r160f16tVNEKXx5Of6V6xYQd++fbGzswP0+/Wai/xcf+3atUlMTAQgMTERa2trLCyKbamx\nLNq1a/fUWjvGyHtFJvHndDNXdHR0nm3MJfnl5/of99NPP/H6668XRmhGl99/+02bNhnKephTMbb8\nXP/ly5eJi4ujY8eOtGzZkqVLlxZ2mEaTn+v38/Pj7Nmz2NjY4OzszHfffVfYYZqMMfJekfmVWdA3\nfBU3z3Id+/fv5+eff+bQoUNGjKjw5OfaJ0yYwFdffWUoM/vkz0Fxlp/rT01NJSQkhH379pGSkkKb\nNm1o3bo1DRo0KIQIjSs/1//FF1/g4uJCUFAQERERdOnShbCwMCwtLQshQtMr6LxXZBK/ra0tUVFR\nhudRUVGGP2tza3P9+nVsbW0LLUZjys/1A5w6dQo/Pz927tz5zKVYi6r8XHtwcDA+Pj6AfqJvx44d\nlClThp49exZqrMaQn+u3t7enWrVqlC9fnvLly9O+fXvCwsLMIvHn5/oPHz7MtGnTAKhfvz5169bl\n4sWLtGzZslBjNQWj5L1/PEtQQFJTU1W9evVUZGSkevToUZ6Tu0eOHDGbyU2l8nf9V69eVfXr11dH\njhwxUZTGkZ9rf5yvr69av359IUZoXPm5/vPnz6tOnTqptLQ0lZycrJo1a6bOnj1roogLVn6u/733\n3lMBAQFKKaVu3bqlbG1t1d27d00RrlFERkbma3K3oPJekenxl/QbvvJz/TNmzCA+Pt4wzl2mTBmO\nHz9uyrALRH6u3Zzl5/obNWqEt7c3zZs3p1SpUvj5+dGkSRMTR14w8nP9H330Ee+88w7Ozs5kZGTw\n9ddfU7VqVRNHXjAGDhzIgQMHiI2Nxd7enk8//ZTU1FTAeHlPbuASQogSpsis6hFCCFE4JPELIUQJ\nI4lfCCFKGEn8QghRwkjiF0KIEkYSvxBClDCS+EWRUbp0aVxdXQ2Pa9eu5dq2UqVK//h8vr6+1KtX\nD1dXV15++WWOHj36zMfw8/PjwoULgL6swOM8PDz+cYzwf9+X5s2b06dPH+7fv//U9mFhYezYsaNA\nzi3Mk6zjF0WGpaUlSUlJBd42N++88w49evSgT58+7Nmzh8mTJxMWFvbcxyuImPI6rq+vL05OTkya\nNCnX9osXLyY4OJjvv/++wGMR5kF6/KLISk5OpnPnzrz88ss0b96czZs3Z2tz8+ZN2rdvj6urK05O\nTvz5558A7N69m7Zt2/Lyyy8zYMAAkpOTczxHZr+nXbt2hr2d586di5OTE05OToYqkMnJyXTr1g0X\nFxecnJxYu3YtAJ6engQHB/PBBx/w4MEDXF1dGTx4MPB/f5X4+Piwfft2wzl9fX3ZsGEDGRkZTJky\nBTc3N5ydnQkMDMzze9KmTRsiIiIAfTnjtm3b0qJFCzw8PLh06RI6nY5PPvmE1atX4+rqytq1a0lO\nTmbYsGG4u7vTokWLHL+PooT5x0UfhCggpUuXVi4uLsrFxUX16dNHpaWlqcTERKWUfvMNR0dHQ9tK\nlSoppZSaPXu2mjlzplJKqfT0dJWUlKRiYmJU+/btVUpKilJKqa+++krNmDEj2/l8fX0NG7qsWbNG\ntW7dWgUHBysnJyeVkpKi7t+/r5o2baq0Wq1at26d8vPzM3w2ISFBKaWUp6enCg4OzhLTkzH+9ttv\naujQoUoppR49eqTs7e3Vw4cP1aJFi9Tnn3+ulFLq4cOHqmXLlioyMjJbnJnHSUtLU3369FE//PCD\nUkqpxMRElZaWppRSas+ePapv375KKaUWL16sxo0bZ/j8hx9+qJYtW6aU0m/o8q9//UslJyfn+G8g\nSoYiU6tHiPLly2fZVi41NZUPP/yQgwcPUqpUKW7cuMGdO3eoUaOGoY2bmxvDhg0jNTWVXr164ezs\nTFBQEOfOnaNt27YA6HQ6w9ePU0oxZcoUPv/8c2rUqMFPP/3Enj176NOnD+XLlwegT58+HDx4EG9v\nbyZPnswHH3xA9+7deeWVV/J9Xd7e3owfPx6dTseOHTvo0KED5cqVY/fu3Zw+fZp169YB+g1GwsPD\ncXBwyPL5zL8koqOjcXBwYNSoUQDcu3ePIUOGEB4ejkajIS0tzXBd6rER3N27d7NlyxZmz54NwKNH\nj4iKiqJhw4b5vgZhXiTxiyJr+fLlxMbGEhISQunSpalbty4PHz7M0qZdu3YcPHiQrVu34uvry8SJ\nE7GysqJLly6sWLHiqcfXaDTMnj2bPn36GF7bu3dvlqSplEKj0dCgQQO0Wi3btm3j448/plOnTkyf\nPj1f1/HCCy/g6enJrl27WLNmDQMHDjS8N3/+fLp06fLUz2f+Qnzw4AFeXl5s2rSJ3r17M336dDp1\n6sRvv/3G1atXn7oP74YNG8yihLMoGDLGL4qsxMREatSoQenSpdm/fz9Xr17N1ubatWtUr16dESNG\nMGLECLRaLa1bt+bQoUOGsfDk5GQuX76c4znUE2sb2rVrx8aNG3nw4AHJycls3LiRdu3acfPmTV54\n4QXefvttJk+enOOG12XKlDH0up/05ptv8vPPPxv+egDw8vJiwYIFhs9cunSJlJSUXL8f5cuXZ968\neUybNg2lFImJidjY2ABkqdhYuXLlLJPMXl5ezJs3z/C8QDbrFsWaJH5RZDy5q9Dbb7/NyZMnad68\nOUuXLqVx48bZ2u7fvx8XFxdatGjBmjVrGD9+PNWqVWPx4sUMHDgQZ2dn2rZty8WLF/N1TldXV3x9\nfXFzc6N169b4+fnh7OzM6dOncXd3x9XVlRkzZvDxxx9nO9a///1vmjdvbpjcffzYr732Gn/88Qdd\nunQx7BU7YsQImjRpQosWLXBycmL06NE5/uJ4/DguLi44OjqyZs0a3n//fT788ENatGhBenq6oV3H\njh05d+6cYXJ3+vTppKam0rx5c5o1a4a/v3/u/wiiRJDlnEIIUcJIj18IIUoYSfxCCFHCSOIXQogS\nRmLPMAsAAAAmSURBVBK/EEKUMJL4hRCihJHEL4QQJYwkfiGEKGEk8QshRAnz/wFPa6W3zlHDwAAA\nAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10a16fc90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEZCAYAAACQK04eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnxBCdMz/A7J8TlzdDymkAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x104547090>"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf = RandomForestClassifier()\n",
      "clf.fit(train[features], train.serious_dlqin2yrs)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 19,
       "text": [
        "RandomForestClassifier(bootstrap=True, compute_importances=None,\n",
        "            criterion='gini', max_depth=None, max_features='auto',\n",
        "            min_density=None, min_samples_leaf=1, min_samples_split=2,\n",
        "            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,\n",
        "            verbose=0)"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "probs = clf.predict_proba(test[features])[::,1]\n",
      "plot_roc(\"RandomForest\", probs)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEVCAYAAADpbDJPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRQ0AIYSYqP8HPZ0ylh12rFQAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10b9cbb90>"
       ]
      }
     ],
     "prompt_number": 20
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###Build your own classifier\n",
      "Pick a different algorithm and a new set of features. Can you beat the 0.74 AUC?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "train.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>serious_dlqin2yrs</th>\n",
        "      <th>revolving_utilization_of_unsecured_lines</th>\n",
        "      <th>age</th>\n",
        "      <th>number_of_time30-59_days_past_due_not_worse</th>\n",
        "      <th>debt_ratio</th>\n",
        "      <th>monthly_income</th>\n",
        "      <th>number_of_open_credit_lines_and_loans</th>\n",
        "      <th>number_of_times90_days_late</th>\n",
        "      <th>number_real_estate_loans_or_lines</th>\n",
        "      <th>number_of_time60-89_days_past_due_not_worse</th>\n",
        "      <th>number_of_dependents</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
        "      <td> 1</td>\n",
        "      <td> 0.766127</td>\n",
        "      <td> 45</td>\n",
        "      <td> 2</td>\n",
        "      <td> 0.802982</td>\n",
        "      <td>  9120</td>\n",
        "      <td> 13</td>\n",
        "      <td> 0</td>\n",
        "      <td> 6</td>\n",
        "      <td> 0</td>\n",
        "      <td> 2</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.957151</td>\n",
        "      <td> 40</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0.121876</td>\n",
        "      <td>  2600</td>\n",
        "      <td>  4</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.658180</td>\n",
        "      <td> 38</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0.085113</td>\n",
        "      <td>  3042</td>\n",
        "      <td>  2</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.907239</td>\n",
        "      <td> 49</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0.024926</td>\n",
        "      <td> 63588</td>\n",
        "      <td>  7</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
        "      <td> 0</td>\n",
        "      <td> 0.213179</td>\n",
        "      <td> 74</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0.375607</td>\n",
        "      <td>  3500</td>\n",
        "      <td>  3</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 15,
       "text": [
        "   serious_dlqin2yrs  revolving_utilization_of_unsecured_lines  age  \\\n",
        "0                  1                                  0.766127   45   \n",
        "1                  0                                  0.957151   40   \n",
        "2                  0                                  0.658180   38   \n",
        "3                  0                                  0.907239   49   \n",
        "4                  0                                  0.213179   74   \n",
        "\n",
        "   number_of_time30-59_days_past_due_not_worse  debt_ratio  monthly_income  \\\n",
        "0                                            2    0.802982            9120   \n",
        "1                                            0    0.121876            2600   \n",
        "2                                            1    0.085113            3042   \n",
        "3                                            1    0.024926           63588   \n",
        "4                                            0    0.375607            3500   \n",
        "\n",
        "   number_of_open_credit_lines_and_loans  number_of_times90_days_late  \\\n",
        "0                                     13                            0   \n",
        "1                                      4                            0   \n",
        "2                                      2                            1   \n",
        "3                                      7                            0   \n",
        "4                                      3                            0   \n",
        "\n",
        "   number_real_estate_loans_or_lines  \\\n",
        "0                                  6   \n",
        "1                                  0   \n",
        "2                                  0   \n",
        "3                                  1   \n",
        "4                                  1   \n",
        "\n",
        "   number_of_time60-89_days_past_due_not_worse  number_of_dependents  \n",
        "0                                            0                     2  \n",
        "1                                            0                     1  \n",
        "2                                            0                     0  \n",
        "3                                            0                     0  \n",
        "4                                            0                     1  "
       ]
      }
     ],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features = ['revolving_utilization_of_unsecured_lines', 'debt_ratio',\n",
      "            'number_of_times90_days_late', 'number_real_estate_loans_or_lines']\n",
      "clf = GradientBoostingClassifier()\n",
      "clf.fit(train[features], train.serious_dlqin2yrs)\n",
      "probs = clf.predict_proba(test[features])[::,1]\n",
      "plot_roc(\"Your Classifier\", probs)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEZCAYAAACQK04eAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfkIIUTGU+AkhRMVQ4ieEEBXz/4xsxvsLrsM8AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x109c0f550>"
       ]
      }
     ],
     "prompt_number": 18
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "###Converting to credit score\n",
      "We're going to take the P(delinquent) outputted by the model and convert it to a FICO style score. We calculate the log odds which we then convert into 'points'. in this case, a increase/decrease in 40 points (arbritrary) means a person's riskness has halved/doubled--40/log(2). We're starting with a base score of 340 (arbitrary)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "probs\n",
      "odds = (1 - probs) / probs\n",
      "score = np.log(odds)*(40/np.log(2)) + 340\n",
      "pl.hist(score)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 19,
       "text": [
        "(array([  2.00000000e+00,   1.20000000e+01,   1.98000000e+02,\n",
        "         6.51000000e+02,   1.08900000e+03,   9.92000000e+02,\n",
        "         6.31900000e+03,   4.44100000e+03,   8.81300000e+03,\n",
        "         1.50680000e+04]),\n",
        " array([ 175.37769656,  218.97334868,  262.5690008 ,  306.16465292,\n",
        "        349.76030504,  393.35595716,  436.95160928,  480.5472614 ,\n",
        "        524.14291352,  567.73856564,  611.33421776]),\n",
        " <a list of 10 Patch objects>)"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAEACAYAAABYq7oeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIg4KBxERcVA4iIiIw/8BWo2XjY08VngAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10884dcd0>"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def convert_prob_to_score(p):\n",
      "    \"\"\"\n",
      "    takes a probability and converts it to a score\n",
      "    Example:\n",
      "        convert_prob_to_score(0.1)\n",
      "        466\n",
      "    \"\"\"\n",
      "    odds = (1 - p) / p\n",
      "    scores = np.log(odds)*(40/np.log(2)) + 340\n",
      "    return scores.astype(np.int)\n",
      "\n",
      "convert_prob_to_score(probs)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 20,
       "text": [
        "array([537, 300, 582, ..., 445, 587, 597])"
       ]
      }
     ],
     "prompt_number": 20
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "##We just did the following\n",
      "\n",
      "- Trained a credit classifier\n",
      "- Evaluated the results by both plotting data and generating reports\n",
      "- Converted the model into a credit score"
     ]
    }
   ],
   "metadata": {}
  }
 ]
}