DS-Meetup-talk.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Hyperparameter Tuning & ML Pipelines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "This is the notebook underlying the reveal.js slides that were used for <a href=\"https://www.meetup.com/Richmond-Data-Science-Community-Meetup/events/250460220/\">Richmond Data Science community meetup on May 30th, 2018.</a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## About me:\n",
    "\n",
    "I am Atul Saurav (@twtAtul), \n",
    "- Lead Genworth Financials' Data Engineering Team\n",
    "- MS in Decison Analytics, VCU DAPT class of 2019\n",
    "- Passionate about learning, data and everything around it\n",
    "\n",
    "You can also find me on LinkedIn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## What is this talk about?\n",
    "\n",
    "- Building Machine Learning Models\n",
    "- Python specific\n",
    "- scikit-learn based models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## What is this talk not about?\n",
    "\n",
    "- Data Cleansing\n",
    "- Feature Engineering\n",
    "- Deep Learning \n",
    "- Other exciting stuff that is difficult to cover in 1 meetup!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "\n",
    "># All models are wrong but some are useful\n",
    "\n",
    " \n",
    "<cite class=\"data-cite\">Box, G. E. P. (1979), \"Robustness in the strategy of scientific model building\", in Launer, R. L.; Wilkinson, G. N., Robustness in Statistics, Academic Press, pp. 201–236.</cite>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Reality is complex - \n",
    "\n",
    "<ul>\n",
    "    <li><p class=\"fragment fade-left\">too many factors influence outcome</p></li>\n",
    "    <li><p class=\"fragment fade-left\">factors difficult to measure accurately and objectively</p></li>\n",
    "    <li><p class=\"fragment fade-left\">not all factors may be known</p></li>\n",
    "</ul>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "> ## How do we build <span class=\"fragment highlight-green\">useful</span> models?\n",
    "> ## How do we <span class=\"fragment highlight-green\">minimize</span> our <span class=\"fragment highlight-red\">effort</span> in model building?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Approach\n",
    "- Use toy datasets for illustration and visualization\n",
    "- Use real dataset for demonstrating application efficacy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Scikit-learn API Overview\n",
    "\n",
    "- All methods are implemented as estimators\n",
    "- All estimators have a ```.fit()``` method\n",
    "- All supervised estimators have ```.predict()``` method\n",
    "- All unsupervised estimators have ```.transform()``` method"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<table>\n",
    "<tr style=\"border:None; font-size:20px; padding:10px;\"><th>``model.predict``</th><th>``model.transform``</th></tr>\n",
    "<tr style=\"border:None; font-size:20px; padding:10px;\"><td>Classification</td><td>Preprocessing</td></tr>\n",
    "<tr style=\"border:None; font-size:20px; padding:10px;\"><td>Regression</td><td>Dimensionality Reduction</td></tr>\n",
    "<tr style=\"border:None; font-size:20px; padding:10px;\"><td>Clustering</td><td>Feature Extraction</td></tr>\n",
    "<tr style=\"border:None; font-size:20px; padding:10px;\"><td>&nbsp;</td><td>Feature Selection</td></tr>\n",
    "\n",
    "</table>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Usual High Level Flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.family import SomeModel\n",
    "myModel = SomeModel()\n",
    "myModel.fit(X_train,y_train)\n",
    "\n",
    "# supervised\n",
    "myModel.predict(X_test)\n",
    "myModel.score(X_test, y_test)\n",
    "\n",
    "## unsupervised\n",
    "myModel.transform(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Problem Statement\n",
    "\n",
    "- Classify data into 2 groups based on already observed groupings\n",
    "- Using house price data, predict if the price of the given house will be >= 500K"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "> ### Binary Classification Problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Toy Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "plt.style.use('seaborn-talk')\n",
    "_ = plt.xkcd()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X ~ n_samples x n_features: (100, 2)\n",
      "y ~ n_samples: (100,)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.datasets import make_blobs\n",
    "\n",
    "scaled_X, scaled_y = make_blobs(centers=2, random_state=0)\n",
    "scaled_X[:,0] = 10* scaled_X[:,0] + 3\n",
    "X, y = make_blobs(centers=2, random_state=0)\n",
    "\n",
    "print('X ~ n_samples x n_features:', X.shape)\n",
    "print('y ~ n_samples:', y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualize Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'first feature')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax0, ax1) = plt.subplots(nrows=1, ncols=2, figsize=(10,5))\n",
    "\n",
    "ax0.scatter(X[y == 0, 0], X[y == 0, 1], c='blue', s=40, label='0')\n",
    "ax0.scatter(X[y == 1, 0], X[y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "ax0.set_xlabel('first feature')\n",
    "ax0.set_ylabel('second feature')\n",
    "\n",
    "ax1.scatter(scaled_X[scaled_y == 0, 0], scaled_X[scaled_y == 0, 1], c='blue', s=40, label='0')\n",
    "ax1.scatter(scaled_X[scaled_y == 1, 0], scaled_X[scaled_y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "ax1.set_xlabel('first feature')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Create Training and Test Set\n",
    "\n",
    "**Training Set** - Data used to train the model\n",
    " \n",
    "**Test Set** - Data held out in the very beginning for testing model performance. This data should not be touched until final scoring"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y,\n",
    "                                                    test_size=0.25,\n",
    "                                                    random_state=1234,\n",
    "                                                    stratify=y)\n",
    "scaled_X_train, scaled_X_test, \\\n",
    "scaled_y_train, scaled_y_test = train_test_split(scaled_X, \n",
    "                                                    scaled_y,\n",
    "                                                    test_size=0.25,\n",
    "                                                    random_state=1234,\n",
    "                                                    stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(75, 2)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(75,)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Train a Nearest Neighbor Classifier - 1 Neighbor\n",
    "\n",
    "Overfit!!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=1, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn = KNeighborsClassifier(n_neighbors=1)\n",
    "\n",
    "knn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=1, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_knn = KNeighborsClassifier(n_neighbors=1)\n",
    "scaled_knn.fit(scaled_X_train, scaled_y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Comapre Model Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.95999999999999996"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_knn.score(scaled_X_test, scaled_y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualize Model - regular"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from figures import plot_2d_separator\n",
    "plt.scatter(X[y == 0, 0], X[y == 0, 1], c='blue', s=40, label='0')\n",
    "plt.scatter(X[y == 1, 0], X[y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "\n",
    "plt.xlabel(\"first feature\")\n",
    "plt.ylabel(\"second feature\")\n",
    "plot_2d_separator(knn, X)\n",
    "_ = plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualize Model - scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from figures import plot_2d_separator\n",
    "plt.scatter(scaled_X[scaled_y == 0, 0], scaled_X[scaled_y == 0, 1], c='blue', s=40, label='0')\n",
    "plt.scatter(scaled_X[scaled_y == 1, 0], scaled_X[scaled_y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "\n",
    "plt.xlabel(\"first feature\")\n",
    "plt.ylabel(\"second feature\")\n",
    "plot_2d_separator(scaled_knn, scaled_X)\n",
    "_ = plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Revelation \n",
    "\n",
    "> ### Scale of various features matters!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Train Nearest Neighbor Classifier - 10 neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=10, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn = KNeighborsClassifier(n_neighbors=10)\n",
    "\n",
    "knn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=10, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_knn = KNeighborsClassifier(n_neighbors=10)\n",
    "scaled_knn.fit(scaled_X_train, scaled_y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Comapre Model Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.83999999999999997"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.80000000000000004"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_knn.score(scaled_X_test, scaled_y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualize Model - regular"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from figures import plot_2d_separator\n",
    "plt.scatter(X[y == 0, 0], X[y == 0, 1], c='blue', s=40, label='0')\n",
    "plt.scatter(X[y == 1, 0], X[y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "\n",
    "plt.xlabel(\"first feature\")\n",
    "plt.ylabel(\"second feature\")\n",
    "plot_2d_separator(knn, X)\n",
    "_ = plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualize Model - scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD+CAYAAAAqP/5ZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzsnXl8G8X5/9+6bUm2LPlOSALNQUISEkhIaChXIBQKJUcJpYVAgZBwlMC3QGm5WqC0UAK00HAVfi1HS6FNCqEFypFynw4tVw6gQDji25ZlWZZkS/P7Y7WbtSLJWlvyOe/Xa1+SVnuMRrOfeWbmmWdMQgiBRI8JkHkiGU2YEq+y3A88Oc17ay4uMsiYej9kSFxTz3B/cPKdP7lG5vfg3Ge453sy+fwfcnFtoRf04faQDmdkXg8sMr8HB5nvA4cAMA92KiQSiUSSG6SgSyQSyQhBCrpEIpGMEKSgSyQSyQhBCrpEIpGMEKSgSyQSyQhBCrpEIpGMEKSgSyQSyQghXzNF3UAE6AacgAUI5OleoExgcAMVQBUwDvABXqAEKE18X5DYihNbAeBIvNoAe+I1uaITQFz3qu6LJV7VzYTyW9XNTOrJFTGUvIkCXYn3XYktBLQCbUAQ8Cde2xP7w4mtI3FMe+K1KfE+wsDM0CtEye8SwIPyP7tQ8tWFkqfuxOZJ7C8BihLnuhPnOBKbDaU8WtLcT+he4+z6P/Sb/jhTYlP/A4vuczJRoJOeeRtJ7A+h5K+6teteG4A6oDHxvp1d5WOgMKPk7RigGqWsF6HkuZrvvsT7Qt2+IpT81+d7pjxSSS5bqcp3HKWMx9iVt2pZbklszYnPrSj5pn6n5m0QpUx3ZpEH+cSEUp5VHfEk3pclPqtl3Zn4XInyXHh039nIrAfJ6Mt4GQa006SL5ZLLWV1hlIdU//ljlD8qgPInhRLv2xPvOxKv+ofQhFIIvSgFUM3QgsR+V2IrS7zmDCEEbW1tNDc309bWRkdHB21tbbS2ttLc3Ex7ezuRSIRoNEo0GqWrq4tQKERHRwednZ1Eo1G6u7uJxWI9rmsymbBYLFitVux2OzabDavVis1mw2az4XQ68fl8FBcXU1RUhMfjweVyUVJSgsfjoaCggIKCAlwuFx6PB5vNlpz0OMrDoQpMK8rD05zIdzWfO1FES02gNbHZUQpnEYroelEEQBVwH7sEI290d3fj9/sJBoN0dHQQCAS0vO3s7CQcDhMMBmlvbycUCmlbNBolEokQDofp6uqiu7tb2+LxOPF4HLXIm0xKkVfzXZ+3DocDm82G2+3G4/Hg8XgoLi6muLhYe19RUYHH49GugyJetUB9YutAKeOqUKmVtFpxxNhV3gvZZVCohoYqwF4UofCh/CclKP+Jm57PmSHa29tpaWmho6ND20KhEO3t7bS3t2v5q75X8zQcDhOJROjq6iIajfYo4yaTSSvbdrudwsJCioqKtE2ffyUlJZSUlGjvvV5vqvLcmci3QCI/mxN5G0Ap52pF4EepAAIo5dqM8iyYE3lqYpfwquW7GKVMVyTyUhXu4sQxPhTN2S1RmYhEIuzcuZPW1lZaWlqor6/Xym84HNbKaiQS0cq0WlZjsRjxeJx9992XG2+8EeBtYE4WtxWQHwvdBNguuOAC3n//fQoLCykpKSnw+XwzVIEqLCzE7Xbj9Xq1P9fn8+Hz+XC5XFituUlWPB6ns7OT9vZ2AoEAoVCIQCBAIBAgGAxSX19PfX09dXV1NDc3a9+1trZSW1tLOBzO/ENNJq3gqoXX5XJRWFiIw+HAYrFgsVgwmUyYTCaEEMRiMSKRCN3d3VpF0N3dTVdXl1Yp+P1+4vHsDL2CggJKSkooLS3F7XbjcrnMPp/PV1ZW5lMflIqKCkpLS3G5XNoDpT5IhYWFekHKCdFolMbGRlpaWjQxaG5uprm5WROGYDBIa2srgUCAtrY22tvbNVEJBoM0NTVlnQcAhYWFFBYWYrfbcTgcFBQUaJWlupnNZm0DpdJWy0h9fb1WUYRCIU28otFoxvva7XYqKiooLy+noqLCXl1dPaGysnJCZWUlTqeTkpISysrK8Hq9lJWVUVJSgtvt1tLQX4QQRCIRzZhQRVk1Rmpra6mrq9Ne6+rqaGlp0f6LbHA4HLjdbgoLC7FarRQUFGgVnt1u18o4QCwWIxwOa4ZOOBzWnr/Ozt6NbafTidvtpqioSM3TwtLS0kKfz1fldDopLy+nrKxMLet4PB68Xq9WOeQiX4UQRKNRQqEQwWCQQCBAY2Mjra2t2mf1N6lGXm1tLY2NjTQ0NNDY2Jjx+haLBafTicPh0PRCX1YtFguhUEg9fJqRtOfDQrcB0QsuuICamhrC4TAtLS34/X7a29t3s1hTXsBmw+FwYLfbcTqdmvXkcDi0H2w2m4nH48RiMe3B6+rq0gRBfSh7w2KxUFFRQUVFhWYRl5SUUFVVRXV1NWVlZZqV7PF48Pl8eL1eiouLsVqtORdDUCoi1VLy+/10dHTg9/tpa2sjHA4TDoe1FoNqZbW0tGjWbHNzMy0tLQQCASKRSK+/3+VyaRWS+tCqLQaz2axVTOrDEovFiMViWqWkpikajRIMBrMSClXsVOu3qKgIp9OJy+WiqKhI+09cLpe2Ty386qY++AUFBTkTyGS6uroIBAL4/X7tQW5ra6OtrY36+noaGhpoaGigqalJE82Ghga6urrSXtNkMmmVqSqKNptNK+OqQJrNZkwmk9ayiEajdHZ2akKjWne9BUw1m81UVFQwZswYqqqqKCsrw+fzMWbMGEpLS7V8d7lcOJ1OrXXodrtxu92prOY+EYvFelTgfr9fy1e/309ra6umE+3t7Vq+NjY24vf79SKXMV9dLpeWr6qOqIJpsVi0tKhlOBKJEIlE6Ozs1FqF2QShtVqtml5UVlZqeTt27FjGjh2rVeSVlZV4PB5Nx2w2mxHdCKO01HpDQH4E3Y3S3N8tEUIIQqEQnZ2dmoXW1tZGIBCgqamJ1tZWzcJQuzPU5onazFObJUIIrftC/1CohVC1lp1Op9bcUy3U4uJi3G435eXllJaW5kWUhwqhUIiGhgYtb1Ux0gtUMBjUxEK1TNVNrTTVPAc0kVeb1mpXhd1ux+124/P5NEtKFQav10t5eTkulyuvAjwUiMfjWhNbbXarLRR9/qtNbdUYUcu4mtfqpoq7w+HoUZmp5Vst6+pntZyXlpZqFaOa3+qzV15ePsi5ZJx4PE5TU5PWutB3g/r9fs1w7Ojo0MqvamioLWG11acvww6HA4fDoRkZbrebgoICTTvUvPT5fLjdbq3Cy0frVsXj8XDWWWexdu3aQRd0gLnAW2PGjOHEE0/kN7/5TQ4vLZFI+sLGjRtZvHgxS5cuZcOGDYOdHEkGKioqWLZsGXfeeachQc+XmRQBpX+3qakpT7eQSCRG6O7uHuwkSLLE7XarXZeGDG1V0HPdbugGpZ80m4EQiUQikexCp52GjG7VnUT1oc4VXaB4HkhBl0gkA4V+7EF1UVW3VKh94KonGqANRuv3DTQ67eyToOcaEyguXb15WUgkEklvqI4SgOYCnE5sVSHWC7O6Pxl1kFQv+tlUBPp7qa96bzD1s/reKDrtHBKC7gQ010KJRCLpC/F4HL/fr85nGZIeaarwq95gqpdSOBzebdKVerzH48no6aXTTkM/OF+CXgRKorLx55RIJJJUhEIh3G43drt9sJOSFn2LIBuffbWS8vl8aY/pq3bmy8vFC2RsFkkkEklvxONxbTLQSEE/WzkdfdXOfAm6G5Q/Qwq6RDI06G2m5VBkpHbb9mZ967TTkJmeL0EvBmmhSyRDge7ubtauXctZZ50FkLGpP9QYqYLeGzrtNCTo+epDLwYlXsJIay5JJMOJjz76iO9///vU1NQAcOqpp6pR/IYFZrM5q/hPIw2ddhqqzfJloZeAEkbS4ehzdE+JRNJPLrvsMmpqahg3bhxPPPEE9913H16vd7CTlTXD0bFCCMHOnTv7dQ2ddqaP8paCvAp6OBymoKAgT7eQSCS9oU7sW7duHcccc8wgp8Y4w63LpaamhgULFjB27FiWLVuWdtyit65onXYamsiTL0H3gRJ6NFehNyUSSd8ZrmNZuRD0TZtgyRKYP1953bQpR4lLIhQKccQRR9DQ0MCSJUsoKCjg4IMP3i2Mdza/R6edmQPyJ5GvPvRyUBY6GMr+oxKJZGijxoPvK5s2wSmnQG3trn1vvgkPPggLF+YggToeeOAB9t57b9544w0tzYsWLWLt2rVcccUV2nFCiF7dFnXaaSh2Sr4s9CKQFrpEIukf/W1Z3HprTzEH5fOtt/brsil58MEH+eEPf9gjDMDEiRN308BsvP902mmoDz1fFnoZKP13hYXZhPKVSCSS3JMs5r3t7w8ff/wxe++9t/Z5x44dbNy4kZdffrnHcdkIuk47h4SFXq6u2lJSUpKnW0gkktFAf6z06mpj+/uD2WzW1n9oampi8eLFXHTRRUyePNnQdZK0M2Dk3HxY6GbAra7L5/HkdXF4iUQywumP2+KaNUqfud4ir65W9uea1atXs2LFCg466CBqampYu3Yt3//+93c7rjcLPUk7/UbSkA9BdwMmv19JhxR0iUQyWCxcqAyAqn3pqpjnekAU4Morr2TOnDk89dRT3HzzzWkt894EPUk724ykIR+CXgloTY/S0tI83EIikUiyY+HC/Ah4MiaTiWOPPZZjjz0243G9CXqSdjYbSUM++tB9AK2trYAUdIlEIjFCknb2WdBzNb/WBbtqmeEUCEgikUgGmyTtbDFybj4sdCfs6gcaTnEjJBLJ0GO4znLNRDZ96AntHBqCrsYwcLlcebiFRCIZLQy34FzZkOk3JWmnoSD2+RD0YoD6+npsNhvFxcV5uIVEIpGMTJK0c9AHRYvURFVUVPQas0AikUgyYTKZRqSVno4k7TQUhzcfalsKUFtbS1VVVR4uL5FIRhsjSdB7q6CStLPeyLXzIehjABoaGqjOx/xaiUQyqhgug6J1dXW0tPQ+htmboOu0sxNoN5KGfEwsqgJobGxk1qxZebi8RCIZTfS5y6WrC957L/V3M2dCjiPBfv3rX6epqYnNmzczZcqUtMf19nt02mlo2j/kx0IvE0LQ0NBARUVFHi4vkUhGE31e5OK992DOnNRbOqHvBwcffDDBYJBzzjkn43GZfk+SdjYZuL0J8jQo2tbWRjQalYIukUhywnDoQz/zzDMB2L59e8bjMlnoSdrZaDQNefFDb2hoAKCysjIPl5dIJNkyHISwN4ZLH/qCBQuorq6mublZm+2ZikyCnqSddUbTkA9BdwQCSghfGWlRIhkcQqEQF198MU899RTAsF6sfbi4LdpsNsaNG0c4HKauLr0WZ/o9SdppuA89H4OiBW1tSsRHKegSycDz4osvcuaZZ/Lxxx9jNpu5+OKLOfzwwwc7WZIsSNLOISHoNrWWKSoqysPlJRJJOjo6Oli0aBHRaJQZM2Zw7733Mm/evMFOVr8YLhZ6LkjSzlaj5+dD0C1qouS0f4lkYOno6CAajVJSUsLmzZvVleNHJzNnwubN6b/LIw6Ho0/nJWmnoeXnIE+CrjYb5HqiEsngYLPZRreYg+Jnvv/+A3rLBx54gEgk0us6oukGepO001AcF8jTmqJqoqSFLpFI+stw8XIBMk4o0pPJbRH6bqHnw8vFFAwGsdvt2HI8E0sikYw+Rkv/OSgLROu0s8Po+XkR9K6uLinmEolEYpAk7YwZPT8vsW0jkciw9nuVSCRDB5vNRjAYJBqNjghrXQhBLJZaq5O0M2r02vnoQ6ejowOn05mPS0skklGG1WrF6/XS3q4EHnQ6nZjNZiwWy5DtX4/FYsRiMeLxeI9N3Zduac4k7Rwagh4Oh6WFLpFIcobFYsHj8RAOh+nu7tYEM9liz+Szroq/2WzGbDZjMpm0LflcIYS2qWKsv67+2OR7mkwmrcJRX202W4/7piNJOw13ueRN0AsLC/NxaYlEMkoxmUz90pVkkdZ/1t8D6CH4qhDrv88XSdqZc0GfDMwAXsZA5K9QKCQFXSKRDCn01rjFYhnk1KQmSTtzOij6XWALsAHYBsxN+t4G7AO4k0+UXi4SiURinCTtNDzdNJ2gm4F1Qgjru+++C+AD/g6cA1wIzEIR+w+AHcC3d7uAXBxaIpFIDKPTTsMDkelUtxQobW5uZv/99+err74C2AO4HbgF+C8wKRG71wf8DdCmSI0E1yKJRCIZaJK0c7fej97IaEar/pIdHcqEpR07dvDWW29p359++uk88MADAHbge0ZvLpFIJJK0lBo9wVC/yNNPP82KFSt61CJbtmxR32rRaEwmU9/WAJRIJP3inXfeGewkSPpBknYaDoaVlaB//PHH2vvt27fzySefaJ9fffVV9e107aJ9XdRVIpH0iWAwyJo1a/jmN78JwOzZswc5RZK+kKSdOetyaQE6y8rKmD59Ou/pVsj2eDyUl5cD4PV69WvnaR34UtAlkoHjww8/ZMaMGdx2222YzWYuv/xyNm7cONjJkvSBJO00vEJQOj/0GLDdZDLN9nq9rFu3jsmTJ/PEE09QUFBAc3MzxcXFuFwu/TlVgAmIW61Wc3d3t9G0SCSSPvDYY4+xY8cOpk2bxp///GdpnQ9jrFYrOu00HD8l08SiN4HZ5557Lt///vdZsWIFNpuNSCTC1772Na0mKSgooKmpibKyshJgLFLQJZIBRbXovv3tb0sxH+YkCbrhLpdMgv4b4NTvfe97BVOnTmXGjBnYbDbC4TBvvPEG0WiUp59+GqfTqV8MehKKoCMFXSKRSIyRpJ2pI3hlOj/Dd1uBHwG377fffmzbto2//OUvnHHGGRx66KEALFq0KPmcvwExm81GV1eX0bRIJBLJqCZJO6uMnt+bl8uJAC+88AJz5szh6quv5rTTTuvhtrh582bmz5/Pa6+9BorfZLygoIBwOGw0LRKJRDKqSdLOPYyen0nQDwUOCwaDnHjiiSxatIg//OEPPP/883z66acA1NTUcMQRRxCNRpkzZ456XtjhcBCJRIymRSKRSEY1SdpZZvT8TF0uqwDuuusuGhoauOSSS7SVrJubm+nq6uLII4+kuLiY9evXqyuM/z/gSLvdXhqNGo7NLpFIJKMau92OTjs9mY5NRSYLfS+A9evXM2vWLBYsWEBhYSGFhYUcccQRfP3rX6e4uJjnn3+er33tawBvARcAHU6nk87OTqNpGXFs2gRLlsD8+crrpk2DnSKJRDKUSdJOJ0pYlazJaoGLoqIiTCYTRUVFzJkzh5dffplx48bpxRygFWWVak3Q4/H4qI26uGkTnHIK1Nbu2vfmm/Dgg7Bw4eClSzL82bQJbr1VKVvV1ZCY5ycZAaTQzmKgqZfTNDKpbQygsLCQL774Qht5ve2227BYLDgcDs477zwuvPBCdbboUcDRJAQdGNUDo+oDp6e2VtkvyQ7Zwtkd1VB47DHFQHjsMXj44cFOlSRXpNBOQ/FcMgn6+wDf+c532LFjB5dddhm33XYbJ598MrFYDJvNxqxZs9i5cyfPPfeces4kwF9UpMxYVRd1HY0ki3lv+yU9SSVcp5wiRT2VoTCKH7MRRwrtNOS6mKnL5T7g7NWrV/Pmm2+ydu1apk2bxvLlyznhhBOYMWNG8vp6AngB2N/tViY4BYNBKisrjaRnxFBdbWy/pCeZWjijuctKGgQjmxTaaahDLZOF/jrwY4vFwh/+8AcCgQAffPABF198MZMnT04W8/8By4F3gU511erRPDC6Zs3u4l1dreyX9I5s4aRGGgQjmxTaacgi7m3E0gFKjF6Hw8E111xDcXExZWVl/PrXv9a715ShRGgE6FQXOR3Ngr5woTIAungxzJunvMoB0ezprYUzWvvXUxkKqt/Bjh0Dnx5JbkmhncaGvEVP0G0HCSFENBoV9957r/ja174mTCaTOOyww8RJJ50kPB6PmDhxonjxxRfVc0NCiIOFEL945plnBCBeeOEFIZH0heeeE6K6WgjYtVVXK/szfTcaeO45IRYsEMJqVX//9QIQLtePR00ejFRSaOda0VOXM20ik4U+H+Cee+7hzDPPZMGCBdTW1vLvf/+bhx56iE8//ZQDDzyQQw45hN///vcAhcBDQFgNq6suXSeRGCVTC2e0exAtXKi4KioxnP4N3A1AR4d11OTBSCWFdhryctEPipqSvnsf4Oyzz+ass87Cau05fur1ennwwQeZNWsWq1at4qCDDmKfffYZC+xTXKykYTR7uYxWkn2k16zpezfTwoWpzx1p/et9ybMvvmgDfowq5jATOGPY5oFEIYV2Ggqhm8lCfwb4pclkiiaJ+VMoYQFeArjkkkuYMGECv/nNb9Tv95IW+uhkoFwNR5IHUV/z7IsvTkQRcxtwDVADTByWeSDZRQrtNLRqUSZBF8DlwG+T9h+NUpIOVnckLWwaU11vpKAPDoM1YDhQXSGpBgYdDti2bfgNkPY1zxyOzxLvngKuBOzSi2oEkEI7DcVzyWbq//cAjjrqKLZv384555zDjBkzAGUk9vbbb+ezzz7ju9/9rnr8+yUlJQeZzWYaGhqMpEXSC6ma5tBz38EHw003DU7IgYHqClH712+9FbZvh08/hUhEeb99+/AKsdDXPEtMKGThwrEEg/3v3pIMDUpKSkjSTkOLXPQm6CUkYvIeccQRdHV1ce211xIKhbQD9thjD/7xj3+oi10EgT9brdbVZWVlUtBzSKrYMC+/rPg4tLTs2vfUU4q46RmoCTkD2RWi9q8vWaJY5nqG0wSk/ubZ7bfD3nvnLj2SwcVqtZKknYa6XHoT9BvUN5deeimXXnop4XCYt956C1Cc4OfOnatOMgoAJwGfgtJ0kIOiuSNV07y5effj0oWhH4jBsjVrFOtYf698dwMM9wHSwcgzydAmSTuzHRQV0PvEooMBjjnmGO69915AEfGDDz6Ygw8+mAMOOACTySSADcBs4EmUiIu4XK6896GPpskl/RWogRgsG4zJVMN9ApKcgCZJJkk7nUbO7c1CtwL873//4/HHH+fMM88EJQrjhYnv48AHwDuAP7FPE3R910yuGW3haY0IssPR01IfSIsvnathvshk4Q6XMjLQeSYZ2iRpp6F46H0JVm4Bbkts64DngUYUzxcnEAViRUVFee1yGW6TS/prKa5ZAzZb78dVV8N1140ei09OQJKMNJK002Lk3KwWuADYunUrNTU12JJUxWQysc8++1itVutZQC3wM6DD4/EUf/nll0bSYojh1HeaC0tx4UI44AB49dXdv6uqgvHje3o6XHRR5vTkavJPrtCnSZ320N2dOn2bNsGVV8KWLcrnffaBa6/d/TcMpzIikah4PB76rJ26MAKpYgN8KIQQkydPFiid7im3E088Ub3Gs4nzvjzjjDPEmDFj8hXyQCxe3DOWh7otXpy3W/aZXKU1FzFMhmIclFRpSk7f2rVKfu29txA22+7H+Hy7/4Z0+V5VJcS8ecr3wzX2yZQpUwQgtm3bNthJkeSYFNqZVRwXIYTI2kL/8Y9/zJVXXrnb/u3bt2shHwHVgSzg9XrH+v3+3Y7PFcPJOyBXlqLe/7qv1vVQjDOeKk16amvh8svTe/CA4rqZ/BtSlRGzGerqlA2GZp96Onbu3ElpaSkOh2OwkyLJI16vl75qZ9Z96C0tLbjd7t22OXPmMH36dFDE/NrE4e1ut5tQKKSfQZpThpN3QC79sxcuhEcfhTfeUF6N/t6B6oYwMmaQzb0ziXm66ySXkaoqSC6O2fSpb9oEBx0EXq+yHXTQwHrLBINBLrzwQvbYYw+mTp3KM888M3A3lww4SdppSECzttB9Pp/69n/AnShuiuOBBUAzsBKoTxzToVrt4XBYWycv1wwX74Ch1JqwpvnHc+nWmGrM4J//VEQ1VV93ru6d6jr6MjJ//i7LXE+mCmXTJli+vOfkrVdfhRNPhEceyX/5e+aZZ1i1ahWfffYZAJ999hlHHXVUfm8qGVSStDNm5NzeLPRWgOuuu46f/OQn6r6JwI3AySh+6hagAsXbRY3v0iHjuexiqLQmNm3afVYlQGlpbiuXVF0o3d2KEKYKPJUqNouebHoYfL7ef0NfWkq33tpTzFWam/PvLXPPPfdw1FFH8dlnnzF79mzeeOMNfvnLX/bocgkGg/lNhGTASdLObiPn9mahfwzMW758OaDEbuns7OT555/nww8/5M033wRg8eLFnHbaaVYUK/0loKm0tBSAxsZGysuNLboxEhkKrYl04rT33rlNW2/94cl93cljA8leLqni09hsitBbrem9XJLpS0upt9+ST1577TUAzjrrLNatW4fNZmPevHksXbqUadOmAfDVV18xZ86c/CZEMqAkaWfOulwuBr4P8P7773PDDTewceNGAoEAADabjSlTprBs2TI1jgvAV4nXBjVRra2tRtIzohlsd8F0AtRtyAbond66UFKlI1OFt2kT7LUXdHaCyQTTpmUn4KnuYXRQOdNvSf7OyP9r5Nj58+f3cBfeuXOq9v6668DtHnxjQZI7krTT2CBkGteYfYQQ8e7ubvHzn/9cmM1mMW7cOHHjjTeKc845RwDi3XffTfa2eUkI4Umcf8Urr7wiAPHUU0/l0qNn2DIU3AUHytVz7VohHI70bohG7jfY+fbcc4pLZPJvKC3tmQYj6cz22DPOOEMA4p577klxruo2vGPQ3U4luSVJO/3CgNtiOkH/gRBC/OlPfxKAuOSSS0Q0GhVCCPH0008LQGzatEmfhqVCCJPu/PPfffddAYi//vWvuf/Fw5Ch4DdvVByfe26X73dVlRBTp/buu52NT7kR8cmUb2r68u1Trq7hWVIihNervM/W5z3V/5vtsakEfde57QK2D+n5F5K+kaSdLSIHfugRgDFjxmAymRg3bpwaUZEjjzySCRMm8Ne//pXDDz9cPf63KN0zlwKfAK2qV0xTU5OhFsNIZSjMWjTS5ZDKU6WuThlUVX23YfdrpfMp93rhkEOMdzOly59t2wYuTks24x9G/t/+lIVdx7iBKYbOlQwPkrQzJ33ozwCthx12mPePf/wjP/jBD7j55puZNGkS8+fP58svv+SFF15gy5ar6D+YAAAgAElEQVQtTJkyBavVOg4YB+wJHAA0qQOhjY2NffpRI42hsmyaKk5qH+5Pf5pa2DNN9qmtVabef/rp7oJanGZJ28mTFb95o6TLn7a23V0QB3OClJH/tz9lYaiUI0n+SNJOQ7Fc0rktNgFLAU499VRqamo488wzCYfD3HTTTcRiMbZs2cL06dOZMGEC06dPJxwOA8xFCdDVarfbcbvdtKRyqxiFpHLNGyxf9GzWsezN4tuyJfWM07a21Mf3VXDS5VtJSerjB8tSNfL/ZnNsJBLh7bffBujhpjiUypEkPyRppyFBV/teUvXJ3JKqfycQCIjGxkZte/nll8Wjjz6qfv2VUPrS9xdCiKqqKrFy5coc9i4Nbwaqz7c3sunDTXeMupWUpN4/dWruBzFT5ZvRMYl85H3yNdV4M9nco7f0nHfeeQIQ48ePFy0tLXn/LZKhhU47gyIHg6IOIURXNBoVN910k4jFYtmkoVMIcXzi/LlCCDF+/Hhx6qmn9uuHSXLPvHmpxXDevF3HZBrcrK5WBgYHc7AyH14l+bq/UVRnBLvdLt58883+X1Ay7NBpZ04EvVwIIRoaGgQgPvzww5Q3feKJJ0RnZ6f68Uzd+bOFUCLC6SIxSoYI2Vq3mbxcBtudUJ++5IojeX+myqev5Mtr6YMPPhBOp1MA4vbbb+/fxSTDFp12hkQOvFySu2UAeOWVV3j77bc5//zzAfjd737H22+/zeWXXw76IfdEv09BQYHaty4ZQmQ7Y7I3747+Rn7sL6nSl8o7J93CIP3pb8+X19KKFSsIhUKcfPLJnH322f27WAYGe5KbJDM67cz9AheqoG/ZsoVf/OIXrFy5ksLCQqDHTNCpulNsoHTuR6NRI+mRDAC5CMOrXmeoiUAq75yurtTH9sczJF/eJv/5z38AuOOOOzRX4VwzXJbmG83otNPQqnLpDo6pF7VYLFpMCYCGhoYeq2k8/fTT6tuJuvOtAFarle5czyuX5IT+huGFobkAczoLOTnKZH89Q/LtbeJyuXJzoRTIpfmGPjrtNCTo6Sz0VqDZ4/GUzp07lw8//BCAeDzO2LFj2WuvvQCYMmWKFtYT0MfEswNYLBZiMUPRHyXDhKFq5aWzkOfNg/Ly3HUx9LWVMxS6OobCJDdJZnTaaaiZlk7QBbAdWFBZWcmNN95IY2MjL7zwAn6/n9///vdMnDiRmpoavQW+B0pXSxc6Qc/XAheSwWUornwE6ccH+hLMC5SY1Lfeeitbt27V9nm9Xv7v//6PhQvHGbrmUKkE01V66WLlSwYenXbmRNBtwNcBfvazn/Hmm2/S0NDA/vvvz6RJk7jooouIxWJa/3hdXR1VVVUFwBhgB5C/9qIkK/JtCebMyuvqgvfeS/3dzJnpRzTTnLuwBB69Cm54YiZf1tt6/+0Z7v9yIMCZq1drLVQ999xzDzfccAOrV6/GbM6uVTxUKsE1a+Dll5WY7nq2bVPKjexHH76kE/Rfk6gZ9t9/f2pTPKWxWIwXXngBp9NJVVUVKKsVqQf6QOmiscpqf8AZCEswZ4OC770H6eJ5b94M++9v+Nx5wPrezs1wjQhK7Oh1JhNCCKbuuScXXnIJ9oQjwMaNG3n00Uc599xzeeihh7jnnnuYMmXK7tdOYqh0dSxcqMTAf/XVnvtTrcuaDUOhG2mk0VftTHfGmQBLly7l3HPP1cc717BYLCzc9a/5gdMB1aWlCBTRlwvaDjwDYQkOpWX1jKIXoIMK4eak7x8EfgdYheAnwBWffYbjwAO1CuIHP/gB69ev57zzzuOll15i1qxZXH311fzoRz/CKkRai3+Pypm8ye4tjsGIw5LOV8Fo5TJUupFGGjrtzMkSdEUAjz/+OE8++WSPm/z0pz9ly5Yt6q4/AAuBrwFP6s4vVI+3WIyFIpD0n4GwBAdiWb229nY+//zz3F2Q5Dg2n/CPF/7LK8ArKGFCAdRFE1ehrHqebJKYTCZOOOEEtm7dymmnnUY4HObSSy9l/vz5vLN+vWLxp9gu/dZ7QyYOS65aWNJjJj/otNOYV4lucpJ+1tF2IYRYsmSJqKysFMFgUAghRH19vQDE2rVr9ZOatgkhHhdCzNSdf50QQsydO1ccffTRuZxAJcmCoRDnJGs2b94toXEQ94HwFhcLs9ksLrnkEhEKhbI6V4BoAdF9//3K90nbsm9HBQQEKLFS9JsJxA9BXJf4fL7+ups3p/0JTz75pBg/frwAhKuwUATTBcDZvDljXn/66adKOkymbMNt9JlczfTNJoyExDg67ewUOZj6/0shhHjuuecEIB566CEhhBB1dXUCEKtWrUqVhhahhAxACHGnEELMmjVLfPvb387Zj5Rkx2DHOTFEkijvAHF0ktACYvLkyeKFF15Ie+4WENeCmJ04vgrE2SCeAREEEUpsC8ffKmB84rpW4WJfsQDE10FYEueaDQq6EErQOq/XKwCxM4OgpyMcDou5c+cKQCxbtiwXOdsruajIh8LCLUIIIaLRlBW42LxZ+W6YodPOgMiBoFcLIUQkEhF77LGHsNvtYs2aNWLx4sUCEBaLRUyZMkXsvffeYurUqeKZZ55Rr7Eqcf5jQgixzz77DFjhlPQk24d10B9InSh/AaI4IaZeEH/8+c/F66+/LqZPn64J+/r163c79+dJ4m9NUSHsvs0R8F+xH7vu/x8Q++uOMSLoQigR8voq6GeffbYAxJ577rlbdMWhzKAbBCppWmvZ/ndDDZ12NoscxHKpBR612+1LHnzwQY477jhu1XWKTZo0ib333htQ+hNLdgWnjiReywG6u7t7LG4rGTh6W3RZ7ftM4ZEHDKDnxcyZijcL8L/NmwmsWsXEPfbg5Xvvperww8Fm4+2332bFihU88sgjPPzww7z++uv885//xOV0svS883h80ybYupXjDz2UVUuXckRxMVvPOIP1wAbgU93tCtwlxEyX0d7+f4CV9+jim2WbufZamFe0lTdOOYWbgXXAkQOUBQ888AB33nkndrudv/3tb3jdbkjEQt+N3lw5B5hchZEYUPrjKjtA6LQzTeCKNOgqhWTF30MI8aW+1ojH4+LFF18U7e3tqSqVT4UQrsS5nwghxLhx48Rpp53W38pKkkN6W/Nz0JrMQohnn31WAOKQQw7Z7btrr722V6tbW5C8F2stbeuln832vljon3zyiSgsLBSAuOuuu7JKvyQFRvJsGOSvTjs/Ezmw0AG+BA4Bfg8cClhMJhMHH3yw+v0W4OPE+49RfNdVB4FygM7OTi2Il2RokGlpOZXB8Lzo6uri6quvBtBCS+jZd999ARg7dizLlixh2b774m9vZ/2mTTz+4ouEwmGmdnenj8SlI23rxWbLzne9N556SokzkMzMmbvtevHFF+ns7OSoo47irLPO6v+9JSMCnXZ2GjmvN8/1T4AjdJ9LUeZtfA58kOacQpQVbIlGo9jtdiPpkeSZdGJeUgJTpgxek/myyy7jpZdeorq6mhtuuGG3748//nj8fj9FRUWY//tfbTLQEpTJD92A87jjtO6bwSSy996w555ZHauGl66urs5bdEXJ8EOnnZHejtVjdCpSMz39zVMxFkAIQUdHB2632+AtJPkknZ/xoYf2bRHnXPD3v/+dtWvXYrFYeOSRR6isrEx5nMfjSbnfntg0dP3yu5HCSs4VHo+Huro6Dj30UO6++26++c1vpj22u7ubW265hauuuko7VzLEGKS+9iTt7OjteD35mJc/BpQmQywWo6ioKA+3kPSVITPDM/GwfPzFF/xgxQoAfr1mDd9wOpXv+vOw5KrrxCB/+9vfOO2003j77bc5+uijOe2007j55pvx+Xw9jtu2bRsrVqygpqYGUBZiV7ub+s0wGPDLGfrf2t2tjM6qTJq067f2tRLvT1iKfpCknQEj5+ZD0KsAAgElHcXFxXm4haSvDBmvhMTD8huUErsU+L9bboFbbsnrw9IvehHLGTNm8MYbb3DTTTfxs5/9jPvuu48nn3ySdevWccIJJ2iHXnjhhdTU1DB+/Hjuuusujj766NylsTcRmjlz5Aj+IAluvknSzqCRc/WCLkgfqtEMnAQcD+yD0lH/DvA88Fd6utaUAfj9fgC9S6NkiDCUVhpqT7wuJkXhSyegulC2WZMLyzULAbFarVx66aUsXbqUlStX8tJLL7F8+XKWLVvGueeei8ViYefOnQDcd999HHbYYanTk68uoxEqgoYYpC65bEnSztaMByeRjYU+AcWdN/mfngecBdwF/Ai4B6VS8AK0tbUBsm9Q0g8yiU8ur5UHIZsyZQrPP/88d911Fz/+8Y/ZsGEDGzZs6HGMzWQaNv7mwMjpzhmkLrlsSdLONiPn9ibopcCLwPiWlhZuuukm/v3vf1NdXc20adP4wQ9+wKRJk9zA3cCBKALvhl3NBinokrzw4IMwbVrPfarbYiqRHISlEM1mM+eccw7HHnssV111lX51LyZMmMABBQXDy1qW1v2AkKSdOe1DvxsY/9prr3Hssccyffp01qxZw6xZs1i/fj3f/e53mTlzJuvWrcPlcp2B4kEWBejoUAZn87k2omSIMBiW27RpuwvI22+nFxz9gNkAM378eP74xz/u/kU66zxf9KWrSjLgJGlnn/vQk6kClnV0dHDWWWdx+umnc8MNN2hB16+44gouv/xybrnlFpYsWcKjjz6Ky+U6G3gWoDmxHIrX6zX6eyTDDWm5DQ9OOWWwUzC8GKS+9iTt9Bs5N5OgLwTYsmULjY2N/OIXv1DF/DHgZeAgk8m05Ec/+hFffPEFN9xwA9dccw3ALICGhgaAtD7FklGI3pLv6lKs5jvvVNZDu+oqJag6DImBqWGLXoS2bjUm4t3dmfv0M5Fs/Q9En3q+BXeQ+tqTtLPeyLmZBH0cwL///W+mTJmin8JvQQkFoLF8+XKOO+44VdA1LxeHwyGn/kt2kcmSN5uHviU/xL0jgOxFKNUYRFdX5pZWJpIrjoFomQ3xwc2+kqSdOfNyCYNSS2zZsoVwOExBQQHAcckHLliwQD9t2QRKx770QZf0i1wK6KRJ/b/WSBKQdGMQ6di6VclDSd5J0s6cDYq+CnDsscdyzjnncMwxx6Sdxh+LxWhpaWHz5s3MSdTwTU1Nu82Qk0h6JV8DrB9/vLtFqr/eYAzsDgeLX+WUU+CNN3qm12iXjiQrkrQzWwvdhLIWQFo+AygrK+PGG2/kuuuu45ZbbmH8+PHEYjGeffZZ9t13X6qrq6mvr6e2tpZoNKqd3NLSQmlpaR9+jmRU058B1lQCqYpOKuHRX28wBnaHm8VvtQ6v9A5TkrSz0ci5mQT9b6AEijnvvPNYtWpVj8UqvvGNb/Q4eMmSJepbAZg6Ojpkl8toYahYmskC2dU18l31MrUspk7N7f8yCL78oxGddnZhcIGLdIJeDBwSDAbZf//9KS0tpaioiPnz51NVVYXZbOaYY45hz54hQj9AGZHdHygJBoOMGTPG8I+RDEOGqqX53nu56RIYDA+ObMnUslAHPnOV3o8+gnnz+n8dSUZ02mnIBx3SC3o3IGw2m2nevHm8//77dHZ2snHjRrZu3UpXYhGB5cuXc//991NQUBAHjgJ2kgjP0dzcLPvQJT1JZclfdRX885+wxx6Dk6ZsGAwPjlygpjvb9M6cCddeC1demd31h0rLbISh005DHi6QXtBDwB8dDsfpDybNsGtqauKJJ57gtNNOY8OGDVx33XVMnjzZDFSiCLoNFNcbKeiSHqSy5PXjLG+/PfK7SIYyNhukWC0q4/HDoWIbZui0M2eCDkpclteBUwBt3Tmr1cq6desYM2YM69atY/LkyaAMoG5RD+nq6iIcDstY6JJdpOjr/bK+nndefz3x4UtYuXIQEibpQSbXROm2mHeStDNk9PxMgh4D1gPL1R3PP/88K1eupLu7m1deeUXfh/57YD5QA1hkpEXJbuj6euMoBeYSlP45j9vN/Bkz0p+r7wvOFQ8+2PN6yd0Ho9UlL1Nf+1AZNxjBJGmnodWKILOgm1A8XQ4LhUKsWLGCDRs2sGTJEn7/+99TVlamP/a6xOsXIANzSTKzDCV+BChx0G//298Yk2pRZZVUk2CSSeXtkan7Ztq0ngKVi+6D4RheNjnNsstrUEnSTkMeLpBZ0PcADmtubmbOnDns2LEDUFZkfzvFjLLq6mpmzpw5DnYtfJuYWSqRaMTYJeYPozT/TJnEPFuy8fbQk87aVwWuL8I2GL7sasuiry0KI/kmBzrzTn+1M5OgmwHi8TgLFixgwYIFCCF45JFHuOuuu3Y7ePny5VqIUCnokt4wAycO9E2zsZKzdQMcKqgti5kzd6+0VPqa3mxaRpKckqSdFqPnZxL0z4G3ysvLD/jzn/+s7YzH44ASvD8dsg9dMqQw6r6XjqEscNLjZETQX+1Mr8rKjM/z1A8fffQRF110EXvttRdz5sxh06ZNycd/DpwLCLmeqMQwatdBqm0oWcQSSR5J0s50azynJZOFXoyyADQvvfQSxxxzDLNnz+byyy+nubmZxYsX8/DDD/Otb30LlPjoqmvjbR0dHRaQg6ISHapgx2LKbEOzGd56a9d30sJMj7qs3scf7/7d5Mmw334DN+A6HAd+hxFJg6JOo+dnEvRzgAlvvfUWxxxzDMcffzz333+/tmLRtGnTOOmkk/jss8/w+XzfAI4AngPibW1tFpAWukSHKtix2K59o0nA9YOsRoXvvffgwAPTfz+QM1fl6lR5Re1ySWhnb0uE7kamEw4DuPvuu9lnn330Yv48cNiSJUs48MAD+ctf/sK5554Lilvxc0C8vb0dQE4skgwc/fX2yGUaYPd06N8PJeGT0/eHFEnamdNBUQ/ABx98wNKlSzXLHPgx8B3g0iuvvJJVq1apgn4oiWn/gUAAs9mM02m4xSCR9I1cddnoBa67WwlIpaJ2fajHJVvZw7HbaDimeQSTpJ05tdDDAFVVVVxzzTXstddenHTSSQBvAneC0u2ybds2vvjiC8aNG1cA7AeYWlpaKCkpyegJIxmdPProo/m9QX8tTr3Avf12emt/KFnZkhFDknYaXr8zk6BvAg6/5ppreO2111izZg1lZWUceeSRAGeDsvjFfvvtx4svvsjJJ58McABAKBSS1rmkB3V1dfzwhz9k/fr1gLISVlYYHYSTFqdkGJOknTkdFL0F+P6MGTOmqS6LixcvZtmyZVx++eVMnTqVL7/8kq1bt+oXsugGJcCMTY52SxJs3bqVb3zjG7S0tOB2u7n++us555xzsjtZDsIZR3qiDFuStNNh9PxMgt4BLAT+5Ha7F951111cdtllXH/99SxatAiHw0FLSwvTp09XXRdjwN+B26SgS/S8/PLLtLS0MGvWLB577DEmTJgw2EkaXsycCa+/nt5tMbkrKZ+VoBxEzStJ2ukjsVZotuf31unuRReTd8KECdxxxx389re/5aWXXqKxsZElS5ZgsVgAXgAaALq7u/WDqBIJAAcccMDoEfNcCp/NBvPnK9tgI7u08kqSdhYAdiCS7fmZVHdv4FVA70zeDfzXbrfPPfzww1m+fDlVVVUcdthhANNJeLlIC10y6pHCJ+kDKbTTRY4E/W6gZMOGDdx333089thjoAj6AUC32Wy2fP3rX2fRokW8++67TJs2rRL4NmCORqPY7XbDP0YiGVLI7gXJAJNCOz1AS7bnpxN0F3BIZ2cnq1evVv3MYZej+wXA7y666CKefPJJnnzySaYpkd72BEyyy0UyIpBWtmSASaGdxemOTUU61e0Cum02m7WoqIh33nlH3W9DEfP7gd+ZTKZUUcHMsVhM7VeXSPqHtJIlo4gU2mnIdTGdoEeBTVar9agVK1ZwzTXXcP311/OTn/wE4DeJjbq6Ov71r3+xYsUK9bwQgBBCTiqS5IbRupJQf5CV4LAlhXbmRNAB1gJHXXHFFXzwwQf89Kc/5fPPP+eKK65gzJgx1NTUsHLlSiZOnMjxxx8PylKRjwF3AJhMhiM/SiT5YbT5ssuuomFNknYaWs4rkxn9DLDBZrPx0EMPsXr1au644w4mTpzI3LlzWbBgAYFAgMcff1xtIqwDatWThcjadVIikUgkCZK009BKF5ks9BXAEgCbzcadd97Jd77zHe6//36EEJxyyimcc845OBwOUNwbL8uQKIlEIpFkQZJ2GlrHM52gW4DfCiHMmzdvZu7cuQAsWrSIRYsWpTr+HiCYeB+3WCzmri7DC1ZLRiBCCLZt2zbYyZBIhgUWi4Uk7TQUgzxdl4sP8DY3NzNv3jy++OKLlAd9+eWX6tu7gX0S77utVisx/UIGklHJJ598wpFHHsnNN98MwPTp0wc5RRLJ0CaFduasDx0hBEIIOjs7AeUBfeWVV7TvV69ezR/+8AdQLH11Efeo3W4nEsl6cpNkBPLwww8zc+ZMNm3aRGlpKX/605+44IILBjtZEsmQJoV2lhk539Dsn+eee45f//rXfPjhh9pI7Icffqh+PSnx2llYWOhWKwHJ6OS2224jFAqxbNky7rzzTsrLDRkauUW68UmGCYWFhSRpp6GFmTMKul60p0yZAsDHH3/MJ598wsSJEwFlAekEMxKvHS6Xq1xd7FQyOonH4wBcdNFFgyvmIN34JMMGl8tFknYaGhRN1+XSAoRKS0uZOXMm77//vvaF1+uloqICgNLSUlpbtWCMauzekNPpTK5lJBKJRNILKbSzJN2xqUhnoceA7SaTaT+v18utt97K+PHjeeyxx7DZbHz11VdMnDgRq9Wqd7GpRond22Wz2YhGo0Z/i0QikYxqUminIS+XTF0uXwH7nX/++SxfvpzzzjsPh8NBOBxWA3EB4HA4aGxspLy83AOMJTEoKgVdIpFIjJFCO3Pih+4DjgM44YQT2LJlC1OmTMFisRCNRvnPf/5DNBrl6aefxul04vV6QZn634ZO0IUQMgSARCKRZEkK7cyJoC8CqKmpoaysrIdFbrfbmT9/Plu2bGHSpEksXbpUDfcYBdqBiMPhQAhBd3e3XOhCIpFIsiSFdvZ5UFRvSrsB3nnnHS677LIeU1GFEPzyl79kzpw5nHnmmSxbtkz93obSgR8sKlK6fQKBQF9+k0QikYxKUminIYs4nZfLlwCHHXYYf/3rX/n73/+ufXH77bdra4p+9NFHfPTRR2zatAmUcAEnAs2lpaUAeg8YiUQikfRCCu3MiaA/C9ROnDiR888/n/PPP59t27bx+eefc9FFF3HzzTczd+5c9tprL8466yz+85//qOfNBloTfeq0tGS9cpJEIpGMelJoZ7aCLiC9oMdQViZi7dq1zJkzh9mzZ3PCCSdw1FFH8b3vfU87sKqqimeffVb9OBXwq6sYtbW1Zf1DJBKJZLSTQjsNrRSU6eC/AneazWYeffRRTj/9dDZv3syqVat6rKhhs9n0jvAWlJmiAMkzniQSiUSSgRTaachNsDf1vwC4xWw2c8cdd9DU1MRxxx0H0ImyohFdXV3MmzdPPb4DaJcWukSd+i+RSLKnv9rZW3CuKPAj4D/AbV6v1wPUAWeg+JxfvHLlSv3x9wERtWO/qampT4mSDF/8fj8XX3wxb7zxBqAEG5JIJNnRX+3MNtriA8AjwATgU0CNwL4SOB+lQ/63wMPAAo/HQ0FBAbW1tamuJRmhbNy4kbPPPpva2lrsdjtXX301s2fPHuxkSSTDhv5qp5HwuRHgw6R99yY2PV+ZTCaqq6upq6vrU6Ikw4+dO3eyZMkShBAsWLCAe++9l6lTpw52siSSYUUK7TS0lqeheOhZ0gKK+43f78/D5SVDkUAggBCCPffckxdffFFdOFwikRgkSTsNDUYZconJkg5AFBcXy0HRUYjD4ZBiLpH0gyTtNLQ4cz4EPQ60FxcX097enofLSyQSycglSTsHXdABWktLS2loaMjT5SUSiWRkkqSdQ0LQd1ZVVdHQ0NAjsJdEIpFIMpOknZHejteTL0FvqKysJBaL0dzcnKdbSCQSycgjSTtjRs7Nl6A3VlZWKm8aG/N0C4lEIhl5JGnnoHu5AHS43W4AgsFgnm4hkUgkI48k7TS0lme+BD1UXFwMyEUuJBKJxAhJ2jkkulyCUtAlEonEOEnaacirJF+C3u10OgEZQlcikUiM0B/tzJegd6m1jJxcJJFIJNmTpJ05jYfeV2LqYqdS0CUSiSR7krSzT4Ju6KQssKpxsEOhUI4vLZFIJCOXJO3sk6DnejqnMJvNFBQUyD50iUQiMUCSdhqKiJuvLhcTKJ37uvVGJRKJRJIFOu00FLo0X4IOKA7ycmKRRCKRGEOnnQ4j5+VL0C2grGAtBV0ikUiModPOISHoDgCbzUZXl6Hoj5JhSixmaEKbRCLJgE47h0SXiwvAbrcTjRoKRSAZhjz//PMsXrwYUCwLiUTSP3TaaTNyXr4EvQikhT7S6ejoYPXq1Rx++OH873//Y+bMmdx7b/Ka4RKJxCg67RwSgu4FsFgssik+grnlllu4++67sdlsXH311dTU1DB79uzBTpZEMuzRaaehLhdDPo4GKAHFn1KuWDRyUQOvXXXVVVxxxRWDnBqJZOTQV+3Ml4VeChCPxzGZcj0JVTLUsNvtg50EiWRE0VftzJege0DxfLBYDLUYJBKJZNSj084hEQ/dB9Dd3Y3Vmq9eHYlEIhmZ6LTTkFdJPgTdRELQI5EIDochv3iJRCIZ9ei0M2LkvHwIupfEYGs4HKagoCAPt5BIJJKRi047B13Qy9Q3oVBIW31DIpFIJNmh005D0Q3zIeil6hsp6BKJRGIcnXYaij+eD0H3qW+i0ah0aZNIJBKD6LTTUOyUfAi6R30jB0UlEonEOENpUNQFittNV1eX7HKRSCQSAyRpp6E1PPMh6G5AW3pORt+TSCSS7EnSzmz70E2QH0EvAmhpaQHA6/Xm4RaSwWb79u3861//ApDhHSSSHJKknX4j5+bNbVFNVFlZWcaDJUgRGnUAABXCSURBVMOLrq4ufvWrXzFr1izeffddqqqq+Na3vjXYyZJIRgxJ2tls5Nx8zMsvh12R+IqLi/NwC8lg8Z3vfIfHH38cgNNPP52bbrpJtsIkkhySpJ1tRs7VC7og0Q/TT7wAbW1KOjweT8aDJcOLV199FYDHHnuM448/fpBTI5GMPJK005Cg56PLxQnQ2toKyD70kcqCBQsGOwkSyYgkSTtbjZybD0F3AOqK1bjd7jzcQiKRSEYmSdrZbuTcvHm5dHYqIQgKCwvzcAuJRCIZmSRp56DHcikCpR/IYrHIiUUSiURigCTtHHQLvRCgvb2doqIi6aMskUgkBkjSzkEfFLWBUsuUlJTk4fISiUQycrnppptoaGhQPwaMnGsSPZeWzoU53UHC02U4rykqhKCtrY3m5mba2tro6Oigra2N1tZWmpubaW9vJxKJEI1GiUajdHV1EQqF6OjooLOzk2g0Snd3N7FYzyUBTSYTFosFq9WK3W7HZrNhtVqx2WzYbDacTic+n4/i4mKKiorweDy4XC5KSkrweDwUFBRQUFCAy+XC4/Fgs9kGNF/Kysrw+/3U1dXlddJYd3c3fr+fYDBIR0cHgUBAy9vOzk7C4TDBYJD29nZCoZC2RaNRIpEI4XCYrq4uuru7tS0ejxOPx7XV1NXWo5rv+rx1OBzYbDbcbjcejwePx0NxcTHFxcXa+4qKCjwez7Bthba3t9PS0kJHR4e2hUIh2tvbaW9v1/JXfa/maTgcJhKJ0NXVRTQa7VHGTSaTVrbtdjuFhYUUFRVpmz7/SkpKKCkp0d57vd4BL8/5IBKJsHPnTlpbW2lpaaG+vl4rv+FwWCurkUhEK9NqWY3FYsTjcfbdd19uvPFGUIIdZivqIh8TiywAF1xwAe+//z6FhYWUlJTg8/k0gSosLMTtduP1erU/1+fz4fP5cLlcOVuHNB6P09nZSXt7O4FAgFAoRCAQIBAIEAwGqa+vp76+nrq6Opqbm7XvWltbqa2tJRwOZ7y+yWTSCq5aeF0uF4WFhTgcDiwWCxaLBZPJhMlkQghBLBYjEonQ3d2tVQRqMB61UvD7/cTj8ax+Y0FBASUlJZSWluJ2u3G5XPh8PsrKyrQHpaKigtLSUlwul/ZAqQ9SYWGhIUFqamrq9ZhoNEpjYyMtLS2aGDQ3N9Pc3KwJQzAYpLW1lUAgQFtbG+3t7ZqoBINBmpqass4DUAaQCgsLsdvtOBwOCgoKtMpS3cxms7aBUmmrZaS+vl6rKEKhkCZe0Wjm6KV2u52KigrKy8upqKigurqayspKKisrcTqdlJSUUFZWhtfrpaysjJKSEtxut5aG/iKEIBKJaMaEKsqqMVJbW0tdXZ32WldXR0tLi/ZfZIPD4cDtdlNYWIjVaqWgoECr8Ox2u1bGQTHiwuGwZuiEw2Ht+VMH+zLhdDpxu90UFRVpeVpaWorP58PpdFJeXk5ZWZlW1j0eD16vV6sccpGvQgii0SihUIhgMEggEKCxsZHW1lbts/qbVCOvtraWxsZGGhoaaGxszHh9tX/c4XBoeqEvqxaLhVBIi8llaFA0HxZ6F2C94IILqKmpIRwO09LSgt/vp729fTeLNRU2mw2Hw4HdbsfpdGrWk8Ph0H6w2WwmHo8Ti8W0B6+rq0sTBPWh7A2LxUJFRQUVFRVahVNSUkJVVRXV1dWUlZVpVrLH48Hn8+H1eikuLsZqtebFOovH45ql5Pf76ejowO/309bWRjgcJhwOay0G1cpqaWnRrNnm5mZaWloIBAJEIpmjb1osFlwul1YhqQ+t2mIwm81axaQ+LLFYjFgsplVKapqi0SjBYDAroVDFTrV+i4qKcDqduFwuioqKtP/E5XJp+9TCr27qg19QUJAzgUymq6uLQCCA3+/XHuS2tjba2tqor6+noaGBhoYGmpqaNNFsaGigqyv92r4mk0mrTFVRtNlsWhlXBdJsNmMymbSWRTQapbOzUxMa1brr+QjvjtlspqKigjFjxlBVVUVZWRk+n48xY8ZQWlqq5bvL5cLpdGqtQ7fbjdvtzpnVHIvFelTgfr9fy1e/309ra6umE+3t7Vq+NjY24vf79SKXMV9dLpeWr6qOqIKp9hjoy3AkEiESidDZ2am1CnvLUwCr1arpRWVlpZa3Y8eOZezYsVpFXllZicfj0XTMZrMZ0Q0r0LtoKoh8CHqMNH3zQghCoRCdnZ2ahdbW1kYgEKCpqYnW1lbNwlC7M9TmidrMU5slQgit+0L/UKiFULWWnU6n1txTLdTi4mLcbjfl5eWUlpamytw4ShziABBEGWkOoQxQNAH1idcOlBq0U/d9A0pAnSBKcPpo4nrqTFwrYAcKUCJTFqL47hcmNh8wFmXlp9LE5yKUGbhFiWPtunPU9xZS/H+hUIiGhgYtb1Ux0gtUMBjUxEK1TNVNrTTVPAc0kVeb1mpXhd1ux+124/P5NEtKFQav10t5eTkul6s3AY4nylA4ke+difcdus+dic8BlIkXbYnNn/gfArpjK4FvAnsm9v0RqNHdz4rSRehNbMWJV1divw8lPlE5SvNX/R/cQEniuIJE/is/IB7Xmthqs1ttoejzX21qq8aIWsbVvFY3VdwdDkePykwt32pZVz+r5by0tFSrGJPyO4ZieKllW81jtZy36/K1BaWsNyc+RxKvwcSxajlXm1NWdpVtVyKPKhP550u8L03s9yXy1JXIU/VZsCXys0ei4/E4TU1NWutC3w3q9/s1w7Gjo0Mrv6qhobaE1Vafvgw7HA4cDodmZLjdbgoKCjTtUPPS5/Phdru1Ci9F61YA3ShlszVRHgPAV4k8bNflebPuuNbEf2AC3knkgznxfVHiutmQF0HX4waqgarEVonycLhRHowylD9U3ZwoAqX/Q62JdGVjggmUgtWt26IohbCDXYW2DSWj64H/JbatQC0GA8oPQQqBqcA4lPyuBiagPERuFMFSBamAXQ+QmV2VgroZIcYuMe5EEYwulPwMoxTqD1AKryoYO1H+gyXAqcDXUP6vx4HTMeiylcRi4EfAIUn7u4EzUMrVVGBvdGKM8uDdBnwErANORMmf14CVwLYM9xybuN4+wB4o+VzBropZFS+1UlbLuNE8F0lbN0qZDqHkdyeKcDSi5PFHie2/QB3ZC8RgY0PJ0yko5dgLTETJz2KUMuxByWcXuyoEVTOM5GucXdoR1b2qxoQfpWwEUCq5FpT8rQe+BLZj0COlF9SKMbt+MYW8C7pkZLAH8CyKWNHU1MQjjzzCW2+9xXnnncfcuXP1x7ahPFQFKa5TA/wY+Hfi8zRgfeI1ebAyBlyRJj0x4FXglRTfWYF7USoIYrEY//znP9m4cSMHHnggK1euzOb3giKMjqQ01QELUSqJTDEtuoEXgTd1+w4HHgTGoFQKJ6MIxiMkfn8SO4DLgacTxxzWS3o/BPbH4BqUkhFFXgZFJSMLG/AEsPcHH3zA9ddfz4YNG+js7KSsrIyTTz4ZUILym81mCgsLPQBfffUV//rXv4jH49hsNpYtW0ZRUdFcYBNwaeKarwOu7du3c+WVV/L444/z61//mvPPPx8U6/VXvaTtXOCOpH2/AE5ta2vjV7/6FY888giffvopXq9XiyvU3d3NP/7xDxoaGnj22Wd7DNadfPLJnHTSSQCOHTt2sHDhwv/f3tkHR1Gle/g3BCZDZuKSQIRQkE9ijFYIooAoFT9W3Xz4AYm3KNRUSFRKQ65GjGgqsO7WglJaAkIKsDYoe/1DwZJQmoDFNZVCWcNNCC4Ly5UIW2IQA64xhHyQMJln/+h0O/mSdXfdBOmn6u2aPn3mnF+f6Xq758x73tGpU6f0wQcf6JZbbpkg6ciPGLvDkrbLcMibJU1sbW2Vx+O5WlKdjCfp0d9995127txp6UhLS9PEiRMjZdwAJEnNzc0qKyvTt99+qz179lgdOBwOvfzyy4qPj79K0i0yxtXmMsV+Qre5GL+U9GF9fb1mzpypc+fOKS8vT4899piuvfZaq9L06dMVEBCg2tparVmzRkuWLOnVSGhoqAoLC7V06dJeoay7d+/WAw88oISEBJWWlio+Pl5HjhzRe++9N6ig0NBQLVq0SDKedP2fbh2SzgHum2++WdXV1UpNTdUjjzyiefPmWfOdd9xxhyorKzVp0iTFxsZq1qxZViRSXl6eJk+erKamJl1//fWaOHGimpqa5PP5VFdXJ4/Ho4MHD2rXrl2D6hs/frxycnL6lXd1dcntdqu2tlbTpk2TJJWXlyszM7NXNI3L5dKjjz6qlStXKjg4WA0NDbrhhht05swZxcfHa+rUqYqJiZEkjRkzRk8//bT5w2WSpD8PKszm5w6iN7LNtj42D2D79u1IYsOGDQzEtGnTiIqK4tNPP0USd955J9XV1dTX11NVVUVGRgYul4vs7Gx8Ph8ABw8eZOzYsTzzzDN0dnZabdXV1TF58mSCgoKYMmVK3/libr31VrPqiT5aAwC8Xi+SSEtLw+v19tOalJTE7bffzoULFwY8F4BXX30Vt9tNe3s7J06cIDAwkJdeegmAvXv3Eh4eTnBwMNHR0f303X333QB8/PHHbNmyxWqzu7ubuLg45s2bB0BraytOp5PExER2795NfX09+/fvJycnh9DQUObMmUNbWxsNDQ2MGDGCF198cVC9wG8G+Oxsu7zM2Pgx1IJsG36WBd879I8++oiBiI2NJSoqis8++wyHw8GJEyf61Vm2bBmSqKqqwuv1Eh0dTWxsLB0dHf3qdnV10dbWxrlz55BEaWkpTU1NNDU10d3dbVbb2EerE7536IWFhQNqvf/++8nIyKCrqwuv14vX67VuMiaZmZkkJydb+3PnziUhIcHqu7Ozk/b2dr7++msk8dZbb1n6zLbmz5/PhAkTerVbWVmJJGpqaujo6OCKK66goqKin8a3334bSaxfvx4At9vNa6+9Zuntc6NaxtBcG7YNLzM2fgy1INuGn60FWL16NZKIiYkhPT2duXPncvLkSUycTicJCQmA8QRcXFzMmTNn2LhxIxkZGaSnpxMVFUVQUBDHjx8HoKCgALfbzezZszl69Ch+1JkvTOe8b98+/+NtwGrA00frbQDHjh1DEm63m7S0NNLT03n//fetN2dmZvZ7ql60aJF1vLOzk0mTJrF06VJrv7y8HEmcPXvWXwfNzc1I4vDhw/QlJCSEBQsWmLtnwLhRRUREEBYWRmdnJw899BD33XcfHR0dbN68maysLNLT00lMTCQgIIBdu3YBEBQU1E9zWVmZ2fbSf+Pnbdula8bGj6EWZNvwsxKAtWvXWo4kMjKSpKQk6uvrMXE6neTn5wOwYsUKwsPDSUhIQBIej4fCwkJeeOEFTp8+jT9NTU0sWLCAkSNH+juovwDBMKBDTwFGDqL1LoBDhw4xevRoJBEaGso111zD66+/bvU5fvx4srOzqaiooKioiHfffZfW1lbreGNjI5JITEzk+eefJzIy0jr3ujrrXvMYDOjQnwWsqafy8nKzfJn5YuHChUiitbWVHTt24HK5uOmmm5CEw+EgLy+P5cuX8/nnnwNw4MABJLFp0ybeeOMNiouLqa6u9v9Wcf0QXRu2DS8zNn4MtSDbhp/NALhw4QKSyM3N9Z/yAODo0aNI4oknnui1bzrVe+65hz78CdiA8aSNz+fjzTffZNy4cXz55ZdmnSwY0KHP+gGtgcBZgNTUVCIiImhpaenbN06nk8rKyn7lJqZDl8SYMWPIysriySefRJL/nPg6GNChLwaoqqpCEjt37jTLi83ziYmJsebZW1paCAkJweVyIYno6Oh+eiorKxk1alS/aSE/1v3AmNh2+ZgdtmhzUQ5I3yeyGmDVofbu3dtrPy4uTlOnTlVKSopaW1u1YcMGLVmyRK+88orZTlKPyWz7wQcfVHFxsbZs2aLly5dL0s3/hNZOSf8n6U6Hw2GlBujBK2nkoUOH1NXVpZqaGo0dO1bHjh2z3nzddddZ0SOStGPHDqWkpCgwMFA+n0/btm3Thx9+qOzsbEmK+0cEgRVEtkKSSktL1djYqD/+0QihDw4OVmpqqpqbm5WWlqb8/Hzde++92rZtm1wuI5S/rq5OXq9XtbW11pJ5k5SUFLnd7v+WtErGIiKbyxjbodv8s6ySsUrv8f37jZX0N954oyTDQY8bN05BQUFatWqVJGnNmjXq7u7W2rVr5XA41NLSIkCjRo3S2bNntXHjRjU0NGjWrFlm+2GSLpq/YwAGSqJyRNLTknZ99dVXkqSioiIVFRVJkpWfp6ysTGafHo9HycnJCgwMlGQsFU9OTvbPDzRtEH2NkpSUlCSn06k9e/YoLS1NknT48GE999xzWrZsmSZMmGC9ISwsTF1dXVq8eLEkKT8/XxkZGdq+fbtcLpdOnjwpwBobh8OhsLAwJSYmKjU11WzmUln9afMTYjt0m4vhk9QxYsSI0XfddZdmzJhhlrfIcOgWppMCdP78eT388MNyOBwqKSmRJK1bt05xcXHKz8/XwoULVVFRoaCgIF155ZW6+uqrtW/fPs2cOVMylln/r6SMmpoauVwuRURE+Pf7Q7RK0pw5cxQbG+v/ngZJmj17tlauXKnbbrvNSjo2ffr0Xjk5KisrFRkZaS5E6pC0UtKKnJwcvfPOO9bpSlJNTY2Cg4MVHh5ultdI+mtISEjM448/rpKSEnk8HgFav369ZsyY0S9G//Tp03rqqackqZdTLygo0KZNm5Sbm6urrrrKGvvQ0FBNmTLFv4mNMtJW2Fzu0JuhngOybXjaq/SmDYg2y0tKSsjNzeX8+fMMhs/nY+vWraxevRowYrCPHz/OF1980XduuBG4CZgAtPl8PhobG81jR4ERF9F6xwDdP9xzbPOgAo0olK8BvvnmG7Zu3WqW7wXGAs2Af+z6CqCzj74DgAP4LzDmx+fPn4/L5cLj8fDss8/S3t5u1l1FTwTRQHzyyScUFBT8gFx8wBfA8wz+I7Ftl5cZGz+GWpBtw9MCgV9jhBNW8H1Uxa/ozx+A/wE+BTYB44AcYPBVPNCJ4Qyfo3co4jzgbz11GoAb/kG99wIfAdU9fTt6yh3AXOC3wEMYP7DOAEL8jlueHMNp3u93ro2AFyP+PQB4APiup+5fgUQ/Db83G2lvb+8ba7+up69gjPDLA8A2YArwS+DbPuPz/8By4G5gJjAN4zMZ6uvCtuFldnIum3+ZBZKyZFw7WyX9QQPP58bKyOHilvS6jMx/v+ip+5WMRFUD4ZQxvXFKxg+bPzUOGel2p8hIsPXnPsec6p2RM1BGRsVT6p23OkDSYhn5ZuJ7yqol/U7S4HkDDMZKKpKRFK1CRk4Xe47c5mLYDt3G5j9AkAyH/KP+fcbG5kdiO3QbGxubnwmXZBz6z+Wr56V28/y5jLs0vMf+Uhzn4Tye/lyKYyv9iPHt+4RuY2NjY3OJ8tP8s66NjY2NzX+cvwMPBoZjRPdJmAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from figures import plot_2d_separator\n",
    "plt.scatter(scaled_X[scaled_y == 0, 0], scaled_X[scaled_y == 0, 1], c='blue', s=40, label='0')\n",
    "plt.scatter(scaled_X[scaled_y == 1, 0], scaled_X[scaled_y == 1, 1], c='red', s=40, label='1', marker='s')\n",
    "\n",
    "plt.xlabel(\"first feature\")\n",
    "plt.ylabel(\"second feature\")\n",
    "plot_2d_separator(scaled_knn, scaled_X)\n",
    "_ = plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Revelation \n",
    "\n",
    "> ### Number of neighbors matters as well!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Tunning # of Neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_scores = []\n",
    "test_scores = []\n",
    "n_neighbors = range(1,28)\n",
    "for neighbor in n_neighbors:\n",
    "    knn = KNeighborsClassifier(n_neighbors=neighbor)\n",
    "    knn.fit(X_train, y_train)\n",
    "    train_scores.append(knn.score(X_train, y_train))\n",
    "    test_scores.append(knn.score(X_test, y_test))\n",
    "\n",
    "plt.plot(n_neighbors, train_scores, label='train')\n",
    "plt.plot(n_neighbors, test_scores, label='test')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('# of neighbors')\n",
    "plt.legend();plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "> ### But Wait!! Is that Hyperparameter Tuning?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "> # No"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "> ### Hyperparameter tuning is part of Model building and Test Data <span style=\"color:red\">should not</span> be used in model build"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "> ### Hyperparameter Tuning should be performed using <span style=\"color:blue\">Validation Set</span> - a subset of training set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "# get columns with null - to return sorted in future\n",
    "def null_pct(df):\n",
    "    return {  k:sum(df[k].isnull())/len(df) for k in df.columns}\n",
    "def null_count(df):\n",
    "    return {  k:sum(df[k].isnull()) for k in df.columns}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.base import TransformerMixin, BaseEstimator\n",
    "\n",
    "class CategoricalTransformer(BaseEstimator, TransformerMixin):\n",
    "    \"Converts a set of columns in a DataFrame to categoricals\"\n",
    "    def __init__(self, columns):\n",
    "        self.columns = columns\n",
    "\n",
    "    def fit(self, X, y=None):\n",
    "        'Records the categorical information'\n",
    "        self.cat_map_ = {col: X[col].astype('category').cat\n",
    "                         for col in self.columns}\n",
    "        return self\n",
    "\n",
    "    def transform(self, X, y=None):\n",
    "        X = X.copy()\n",
    "        for col in self.columns:\n",
    "            X[col] = pd.Categorical(X[col],\n",
    "            categories=self.cat_map_[col].categories,\n",
    "            ordered=self.cat_map_[col].ordered)\n",
    "        return X\n",
    "\n",
    "    def inverse_transform(self, trn, y=None):\n",
    "        trn = trn.copy()\n",
    "        trn[self.columns] = trn[self.columns].apply(lambda x: x.astype(object))\n",
    "        return trn\n",
    "    \n",
    "class DummyEncoder(BaseEstimator, TransformerMixin):\n",
    "\n",
    "    def fit(self, X, y=None):\n",
    "        self.columns_ = X.columns\n",
    "        self.cat_cols_ = X.select_dtypes(include=['category']).columns\n",
    "        self.non_cat_cols_ = X.columns.drop(self.cat_cols_)\n",
    "        self.cat_map_ = {col: X[col].cat for col in self.cat_cols_}\n",
    "\n",
    "        self.cat_blocks_ = {}  # {cat col: slice}\n",
    "        left = len(self.non_cat_cols_)\n",
    "        for col in self.cat_cols_:\n",
    "            right = left + len(self.cat_map_[col].categories)\n",
    "            self.cat_blocks_[col] = slice(left, right)\n",
    "            left = right\n",
    "        return self\n",
    "\n",
    "    def transform(self, X, y=None):\n",
    "        return np.asarray(pd.get_dummies(X))\n",
    "\n",
    "    def inverse_transform(self, trn, y=None):\n",
    "        numeric = pd.DataFrame(trn[:, :len(self.non_cat_cols_)],\n",
    "                               columns=self.non_cat_cols_)\n",
    "        series = []\n",
    "        for col, slice_ in self.cat_blocks_.items():\n",
    "            codes = trn[:, slice_].argmax(1)\n",
    "            cat = self.cat_map_[col]\n",
    "            cat = pd.Categorical.from_codes(codes,\n",
    "                                            cat.categories,\n",
    "                                            cat.ordered)\n",
    "            series.append(pd.Series(cat, name=col))\n",
    "        return pd.concat([numeric] + series, axis='columns')[self.columns_]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Real Life Example - Housing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sale_type</th>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLS Listing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sold_date</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>property_type</th>\n",
       "      <td>Condo/Co-op</td>\n",
       "      <td>Single Family Residential</td>\n",
       "      <td>Single Family Residential</td>\n",
       "      <td>Single Family Residential</td>\n",
       "      <td>Single Family Residential</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>city</th>\n",
       "      <td>Kew Gardens</td>\n",
       "      <td>Anaheim</td>\n",
       "      <td>Howard Beach</td>\n",
       "      <td>Aliso Viejo</td>\n",
       "      <td>Orlando</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <td>NY</td>\n",
       "      <td>CA</td>\n",
       "      <td>NY</td>\n",
       "      <td>CA</td>\n",
       "      <td>FL</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zip</th>\n",
       "      <td>11415</td>\n",
       "      <td>92807</td>\n",
       "      <td>11414</td>\n",
       "      <td>92656</td>\n",
       "      <td>32837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beds</th>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>baths</th>\n",
       "      <td>1</td>\n",
       "      <td>5.5</td>\n",
       "      <td>1.5</td>\n",
       "      <td>4.5</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>location</th>\n",
       "      <td>The Texas</td>\n",
       "      <td>91 - Sycamore Canyon</td>\n",
       "      <td>Howard Beach</td>\n",
       "      <td>AV - Aliso Viejo</td>\n",
       "      <td>Orlando</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>square_feet</th>\n",
       "      <td>NaN</td>\n",
       "      <td>7400</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3258</td>\n",
       "      <td>1596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>lot_size</th>\n",
       "      <td>NaN</td>\n",
       "      <td>56628</td>\n",
       "      <td>2400</td>\n",
       "      <td>5893</td>\n",
       "      <td>5623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>year_built</th>\n",
       "      <td>1956</td>\n",
       "      <td>2000</td>\n",
       "      <td>1950</td>\n",
       "      <td>2011</td>\n",
       "      <td>1994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days_on_market</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>15</td>\n",
       "      <td>6</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>x__square_feet</th>\n",
       "      <td>NaN</td>\n",
       "      <td>514</td>\n",
       "      <td>NaN</td>\n",
       "      <td>457</td>\n",
       "      <td>166</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hoa_month</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>258</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>status</th>\n",
       "      <td>Active</td>\n",
       "      <td>Active</td>\n",
       "      <td>Active</td>\n",
       "      <td>Active</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>next_open_house_start_time</th>\n",
       "      <td>March-11-2018 01:00 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>next_open_house_end_time</th>\n",
       "      <td>March-11-2018 03:00 PM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <td>MLSLI</td>\n",
       "      <td>CRMLS</td>\n",
       "      <td>MLSLI</td>\n",
       "      <td>CRMLS</td>\n",
       "      <td>MFRMLS</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>favorite</th>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>interested</th>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>latitude</th>\n",
       "      <td>40.7</td>\n",
       "      <td>33.8</td>\n",
       "      <td>40.7</td>\n",
       "      <td>33.6</td>\n",
       "      <td>28.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>longitude</th>\n",
       "      <td>-73.8</td>\n",
       "      <td>-117.8</td>\n",
       "      <td>-73.8</td>\n",
       "      <td>-117.7</td>\n",
       "      <td>-81.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 0                          1  \\\n",
       "id                                               1                          2   \n",
       "sale_type                              MLS Listing                MLS Listing   \n",
       "sold_date                                      NaN                        NaN   \n",
       "property_type                          Condo/Co-op  Single Family Residential   \n",
       "city                                   Kew Gardens                    Anaheim   \n",
       "state                                           NY                         CA   \n",
       "zip                                          11415                      92807   \n",
       "beds                                             0                          7   \n",
       "baths                                            1                        5.5   \n",
       "location                                 The Texas       91 - Sycamore Canyon   \n",
       "square_feet                                    NaN                       7400   \n",
       "lot_size                                       NaN                      56628   \n",
       "year_built                                    1956                       2000   \n",
       "days_on_market                                   1                          2   \n",
       "x__square_feet                                 NaN                        514   \n",
       "hoa_month                                      NaN                        NaN   \n",
       "status                                      Active                     Active   \n",
       "next_open_house_start_time  March-11-2018 01:00 PM                        NaN   \n",
       "next_open_house_end_time    March-11-2018 03:00 PM                        NaN   \n",
       "source                                       MLSLI                      CRMLS   \n",
       "favorite                                         N                          N   \n",
       "interested                                       Y                          Y   \n",
       "latitude                                      40.7                       33.8   \n",
       "longitude                                    -73.8                     -117.8   \n",
       "target                                       False                       True   \n",
       "\n",
       "                                                    2  \\\n",
       "id                                                  3   \n",
       "sale_type                                 MLS Listing   \n",
       "sold_date                                         NaN   \n",
       "property_type               Single Family Residential   \n",
       "city                                     Howard Beach   \n",
       "state                                              NY   \n",
       "zip                                             11414   \n",
       "beds                                                3   \n",
       "baths                                             1.5   \n",
       "location                                 Howard Beach   \n",
       "square_feet                                       NaN   \n",
       "lot_size                                         2400   \n",
       "year_built                                       1950   \n",
       "days_on_market                                     15   \n",
       "x__square_feet                                    NaN   \n",
       "hoa_month                                         NaN   \n",
       "status                                         Active   \n",
       "next_open_house_start_time                        NaN   \n",
       "next_open_house_end_time                          NaN   \n",
       "source                                          MLSLI   \n",
       "favorite                                            N   \n",
       "interested                                          Y   \n",
       "latitude                                         40.7   \n",
       "longitude                                       -73.8   \n",
       "target                                           True   \n",
       "\n",
       "                                                    3  \\\n",
       "id                                                  4   \n",
       "sale_type                                 MLS Listing   \n",
       "sold_date                                         NaN   \n",
       "property_type               Single Family Residential   \n",
       "city                                      Aliso Viejo   \n",
       "state                                              CA   \n",
       "zip                                             92656   \n",
       "beds                                                4   \n",
       "baths                                             4.5   \n",
       "location                             AV - Aliso Viejo   \n",
       "square_feet                                      3258   \n",
       "lot_size                                         5893   \n",
       "year_built                                       2011   \n",
       "days_on_market                                      6   \n",
       "x__square_feet                                    457   \n",
       "hoa_month                                         258   \n",
       "status                                         Active   \n",
       "next_open_house_start_time                        NaN   \n",
       "next_open_house_end_time                          NaN   \n",
       "source                                          CRMLS   \n",
       "favorite                                            N   \n",
       "interested                                          Y   \n",
       "latitude                                         33.6   \n",
       "longitude                                      -117.7   \n",
       "target                                           True   \n",
       "\n",
       "                                                    4  \n",
       "id                                                  5  \n",
       "sale_type                                 MLS Listing  \n",
       "sold_date                                         NaN  \n",
       "property_type               Single Family Residential  \n",
       "city                                          Orlando  \n",
       "state                                              FL  \n",
       "zip                                             32837  \n",
       "beds                                                3  \n",
       "baths                                               2  \n",
       "location                                      Orlando  \n",
       "square_feet                                      1596  \n",
       "lot_size                                         5623  \n",
       "year_built                                       1994  \n",
       "days_on_market                                      8  \n",
       "x__square_feet                                    166  \n",
       "hoa_month                                          64  \n",
       "status                                         Active  \n",
       "next_open_house_start_time                        NaN  \n",
       "next_open_house_end_time                          NaN  \n",
       "source                                         MFRMLS  \n",
       "favorite                                            N  \n",
       "interested                                          Y  \n",
       "latitude                                         28.4  \n",
       "longitude                                       -81.4  \n",
       "target                                          False  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('new_train.csv')\n",
    "data.columns = [ x.lower().replace('.','_') for x in data.columns]\n",
    "data.head().T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Some Observations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 19318 entries, 0 to 19317\n",
      "Data columns (total 25 columns):\n",
      "id                            19318 non-null int64\n",
      "sale_type                     19318 non-null object\n",
      "sold_date                     0 non-null float64\n",
      "property_type                 19318 non-null object\n",
      "city                          19306 non-null object\n",
      "state                         19318 non-null object\n",
      "zip                           19271 non-null object\n",
      "beds                          18216 non-null float64\n",
      "baths                         18053 non-null float64\n",
      "location                      18773 non-null object\n",
      "square_feet                   15693 non-null float64\n",
      "lot_size                      10267 non-null float64\n",
      "year_built                    16950 non-null float64\n",
      "days_on_market                18328 non-null float64\n",
      "x__square_feet                15693 non-null float64\n",
      "hoa_month                     7553 non-null float64\n",
      "status                        19318 non-null object\n",
      "next_open_house_start_time    933 non-null object\n",
      "next_open_house_end_time      933 non-null object\n",
      "source                        19318 non-null object\n",
      "favorite                      19318 non-null object\n",
      "interested                    19318 non-null object\n",
      "latitude                      19318 non-null float64\n",
      "longitude                     19318 non-null float64\n",
      "target                        19318 non-null bool\n",
      "dtypes: bool(1), float64(11), int64(1), object(12)\n",
      "memory usage: 3.6+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### What's different?\n",
    "\n",
    "- Not all features are numeric - **Categorical Variables **\n",
    "- Lot of missing data points"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "### More Revalations:\n",
    "> ### scikit-learn models need data to be numeric/float\n",
    "> ### scikit-learn models implicitly cannot handle missing values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Final Revelation\n",
    "\n",
    "> ### Any transformation applied on the training set to handle first 3 revelations should later be applied on the test set as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "del(data['sold_date'])\n",
    "del(data['next_open_house_start_time'])\n",
    "del(data['next_open_house_end_time'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Summary so far\n",
    "\n",
    "We need to string following steps into a managable fashion to build <span style=\"color:red\">effective</span> models\n",
    "\n",
    "- Handle missing data\n",
    "- Handle different scales across features\n",
    "- Handle categorical data\n",
    "\n",
    "### - Pipelines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "#### But then also tune the model\n",
    "\n",
    "- split training data into (cross)validation sets\n",
    "- search best values for hyperparameters for optimal model performance\n",
    "\n",
    "### - GridSearch"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### More on Pipelines and GridSearchCV\n",
    "\n",
    "- also known as meta-estimators in scikit-learn\n",
    "- they inherit the properties of the last estimator used\n",
    "- pipeline is used to chain multiple estimator into one pipe so output of one flows as input of next\n",
    "- GridSearchCV is used to search on a hyperparameter grid the model with the best score "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Making ML Pipelines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.pipeline import make_pipeline\n",
    "make_pipeline( CategoricalTransformer(columns=cat_cols), DummyClassifier(\"most_frequent\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "This will inherit the properties of DummyClassifier\n",
    "\n",
    "![pipeline](figures/pipeline.png)\n",
    "\n",
    "Image Source: <a href=\"https://github.com/amueller/scipy-2017-sklearn/\">SciPy 2016 Scikit-learn Tutorial</a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "> ## Back to real life housing classification problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Which features have less than 20% missing values?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>null_pct</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>latitude</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>interested</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>favorite</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>status</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>longitude</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>state</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>property_type</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sale_type</th>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>city</th>\n",
       "      <td>0.000621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>zip</th>\n",
       "      <td>0.002433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>location</th>\n",
       "      <td>0.028212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days_on_market</th>\n",
       "      <td>0.051248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beds</th>\n",
       "      <td>0.057045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>baths</th>\n",
       "      <td>0.065483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>year_built</th>\n",
       "      <td>0.122580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>square_feet</th>\n",
       "      <td>0.187649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>x__square_feet</th>\n",
       "      <td>0.187649</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                null_pct\n",
       "id              0.000000\n",
       "latitude        0.000000\n",
       "interested      0.000000\n",
       "favorite        0.000000\n",
       "source          0.000000\n",
       "status          0.000000\n",
       "longitude       0.000000\n",
       "target          0.000000\n",
       "state           0.000000\n",
       "property_type   0.000000\n",
       "sale_type       0.000000\n",
       "city            0.000621\n",
       "zip             0.002433\n",
       "location        0.028212\n",
       "days_on_market  0.051248\n",
       "beds            0.057045\n",
       "baths           0.065483\n",
       "year_built      0.122580\n",
       "square_feet     0.187649\n",
       "x__square_feet  0.187649"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d = pd.DataFrame(null_pct(data), index=['null_pct']).T.sort_values('null_pct')\n",
    "d[(d.null_pct< .2 )]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Set features to work with"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>favorite</th>\n",
       "      <th>interested</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>status</th>\n",
       "      <th>property_type</th>\n",
       "      <th>sale_type</th>\n",
       "      <th>source</th>\n",
       "      <th>state</th>\n",
       "      <th>beds</th>\n",
       "      <th>baths</th>\n",
       "      <th>year_built</th>\n",
       "      <th>x__square_feet</th>\n",
       "      <th>square_feet</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>40.7</td>\n",
       "      <td>-73.8</td>\n",
       "      <td>Active</td>\n",
       "      <td>Condo/Co-op</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLSLI</td>\n",
       "      <td>NY</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1956.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>33.8</td>\n",
       "      <td>-117.8</td>\n",
       "      <td>Active</td>\n",
       "      <td>Single Family Residential</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>CRMLS</td>\n",
       "      <td>CA</td>\n",
       "      <td>7.0</td>\n",
       "      <td>5.5</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>514.0</td>\n",
       "      <td>7400.0</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>40.7</td>\n",
       "      <td>-73.8</td>\n",
       "      <td>Active</td>\n",
       "      <td>Single Family Residential</td>\n",
       "      <td>MLS Listing</td>\n",
       "      <td>MLSLI</td>\n",
       "      <td>NY</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.5</td>\n",
       "      <td>1950.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id favorite interested  latitude  longitude  status  \\\n",
       "0   1        N          Y      40.7      -73.8  Active   \n",
       "1   2        N          Y      33.8     -117.8  Active   \n",
       "2   3        N          Y      40.7      -73.8  Active   \n",
       "\n",
       "               property_type    sale_type source state  beds  baths  \\\n",
       "0                Condo/Co-op  MLS Listing  MLSLI    NY   0.0    1.0   \n",
       "1  Single Family Residential  MLS Listing  CRMLS    CA   7.0    5.5   \n",
       "2  Single Family Residential  MLS Listing  MLSLI    NY   3.0    1.5   \n",
       "\n",
       "   year_built  x__square_feet  square_feet  target  \n",
       "0      1956.0             NaN          NaN   False  \n",
       "1      2000.0           514.0       7400.0    True  \n",
       "2      1950.0             NaN          NaN    True  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_features = ['id', 'favorite', 'interested', 'latitude', 'longitude', 'status', 'property_type', 'sale_type', 'source', \n",
    "                'state', 'beds', 'baths', 'year_built', 'x__square_feet', 'square_feet', 'target']\n",
    "\n",
    "sub_data = data[new_features]\n",
    "sub_data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19318\n",
      "17815\n"
     ]
    }
   ],
   "source": [
    "print (len(sub_data))\n",
    "sub_data = sub_data.drop_duplicates(subset=['favorite', 'interested', 'latitude', 'longitude', 'status', 'property_type', 'sale_type', \n",
    "                                            'source', 'state', 'beds', 'baths', 'year_built', 'x__square_feet', 'square_feet', 'target'])\n",
    "print (len(sub_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "sub_data = sub_data.copy()\n",
    "sub_data.loc[sub_data['sale_type'] == 'New Construction Plan','year_built'] = 2018.0\n",
    "sub_data.loc[sub_data['property_type'] == 'Vacant Land','year_built'] = 2019.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.dummy import DummyClassifier\n",
    "from sklearn.model_selection import cross_val_score, train_test_split\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import roc_auc_score, roc_curve, make_scorer\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.linear_model import LogisticRegressionCV\n",
    "from sklearn.svm import SVC, LinearSVC\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.pipeline import Pipeline, make_pipeline\n",
    "from sklearn.preprocessing import Imputer, RobustScaler, StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "def plot_roc(model, X_test, y_test):\n",
    "    df = model.decision_function(X_test)\n",
    "    fpr, tpr, _ = roc_curve(y_test, df)\n",
    "    acc = model.score(X_test, y_test)\n",
    "    auc0 = roc_auc_score(y_test, df)\n",
    "    auc1 = roc_auc_score(y_test, model.predict(X_test))\n",
    "    plt.plot(fpr, tpr, label=\"acc:%.2f auc0:%.2f auc1:%.2f\" % (acc, auc0, auc1), linewidth=3)\n",
    "    plt.xlabel(\"False Positive Rate\")\n",
    "    plt.ylabel(\"True Positive Rate (recall)\")\n",
    "    plt.title(repr(model).split('(')[0])\n",
    "    plt.legend(loc=\"best\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "roc_auc_scorer = make_scorer(roc_auc_score, greater_is_better=True,\n",
    "                             needs_threshold=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Set model Inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "X = sub_data[['favorite', 'interested', 'latitude', 'longitude', 'status', 'property_type', 'sale_type', \n",
    "                    'source', 'state', 'beds', 'baths', 'year_built', 'x__square_feet', 'square_feet']]\n",
    "y = sub_data.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "cat_cols = ['favorite', 'interested', 'status', 'property_type', 'sale_type', 'source', 'state']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Establish base case for prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.68984678,  0.68996295,  0.68996295])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_pipe = make_pipeline( CategoricalTransformer(columns=cat_cols), DummyEncoder(), DummyClassifier(\"most_frequent\"))\n",
    "cross_val_score(dummy_pipe, X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "### Create training and test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X,y, random_state=42 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "Cs = [ 1, 10, 100, 500, 750] \n",
    "gammas = [0.0005, 0.001, 0.01, .1, 1] \n",
    "param_grid = {'svc__C': Cs, 'svc__gamma' : gammas}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "svc_pipe = make_pipeline( CategoricalTransformer(columns=cat_cols), DummyEncoder(), Imputer(strategy='median'),  StandardScaler(), SVC(random_state=42) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true,
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 25 candidates, totalling 125 fits\n",
      "[CV] svc__C=1, svc__gamma=0.0005 .....................................\n",
      "[CV] svc__C=1, svc__gamma=0.0005 .....................................\n",
      "[CV] svc__C=1, svc__gamma=0.0005 .....................................\n",
      "[CV] svc__C=1, svc__gamma=0.0005 .....................................\n",
      "[CV] svc__C=1, svc__gamma=0.0005 .....................................\n",
      "[CV] svc__C=1, svc__gamma=0.001 ......................................\n",
      "[CV] svc__C=1, svc__gamma=0.001 ......................................\n",
      "[CV] svc__C=1, svc__gamma=0.001 ......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.001, score=0.8323980546202768, total=  18.0s\n",
      "[CV] svc__C=1, svc__gamma=0.001 ......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.0005, score=0.7875046763935653, total=  19.0s\n",
      "[CV] svc__C=1, svc__gamma=0.001 ......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.001, score=0.7639356528245417, total=  19.5s\n",
      "[CV] svc__C=1, svc__gamma=0.01 .......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.001, score=0.780314371257485, total=  19.8s\n",
      "[CV] svc__C=1, svc__gamma=0.01 .......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.0005, score=0.7542087542087542, total=  20.3s\n",
      "[CV] svc__C=1, svc__gamma=0.01 .......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.0005, score=0.7746162485960314, total=  20.6s\n",
      "[CV] svc__C=1, svc__gamma=0.01 .......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.0005, score=0.7784431137724551, total=  21.1s\n",
      "[CV] svc__C=1, svc__gamma=0.01 .......................................\n",
      "[CV]  svc__C=1, svc__gamma=0.0005, score=0.7690868263473054, total=  20.8s\n",
      "[CV] svc__C=1, svc__gamma=0.1 ........................................\n",
      "[CV]  svc__C=1, svc__gamma=0.01, score=0.8787878787878788, total=  14.1s\n",
      "[CV] svc__C=1, svc__gamma=0.1 ........................................\n",
      "[CV]  svc__C=1, svc__gamma=0.01, score=0.844311377245509, total=  17.8s\n",
      "[CV]  svc__C=1, svc__gamma=0.01, score=0.8352676900037439, total=  17.1s\n",
      "[CV]  svc__C=1, svc__gamma=0.01, score=0.8166853722409277, total=  17.8s\n",
      "[CV]  svc__C=1, svc__gamma=0.01, score=0.8334580838323353, total=  17.5s\n",
      "[CV] svc__C=1, svc__gamma=0.1 ........................................\n",
      "[CV] svc__C=1, svc__gamma=0.1 ........................................\n",
      "[CV]  svc__C=1, svc__gamma=0.1, score=0.8327721661054994, total=  16.3s\n",
      "[CV] svc__C=1, svc__gamma=0.1 ........................................\n",
      "[CV] svc__C=1, svc__gamma=1 ..........................................\n",
      "[CV] svc__C=1, svc__gamma=1 ..........................................\n",
      "[CV]  svc__C=1, svc__gamma=0.001, score=0.7907934131736527, total=  20.1s\n",
      "[CV]  svc__C=1, svc__gamma=0.001, score=0.7858479970048671, total=  19.8s\n",
      "[CV] svc__C=1, svc__gamma=1 ..........................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:   59.1s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] svc__C=1, svc__gamma=1 ..........................................\n",
      "[CV]  svc__C=1, svc__gamma=0.1, score=0.8843995510662177, total=  13.9s\n",
      "[CV] svc__C=1, svc__gamma=1 ..........................................\n",
      "[CV] . svc__C=1, svc__gamma=1, score=0.9042274597830153, total=  15.0s\n",
      "[CV] svc__C=10, svc__gamma=0.0005 ....................................\n",
      "[CV]  svc__C=1, svc__gamma=0.1, score=0.847997004867091, total=  15.0s\n",
      "[CV] svc__C=10, svc__gamma=0.0005 ....................................\n",
      "[CV]  svc__C=1, svc__gamma=0.1, score=0.8510479041916168, total=  15.8s\n",
      "[CV] svc__C=10, svc__gamma=0.0005 ....................................\n",
      "[CV]  svc__C=1, svc__gamma=0.1, score=0.8401946107784432, total=  15.7s\n",
      "[CV] svc__C=10, svc__gamma=0.0005 ....................................\n",
      "[CV] . svc__C=1, svc__gamma=1, score=0.8514777403666293, total=  16.5s\n",
      "[CV] svc__C=10, svc__gamma=0.0005 ....................................\n",
      "[CV] . svc__C=1, svc__gamma=1, score=0.8589071856287425, total=  15.5s\n",
      "[CV] svc__C=10, svc__gamma=0.001 .....................................\n",
      "[CV] . svc__C=1, svc__gamma=1, score=0.8581586826347305, total=  16.0s\n",
      "[CV] svc__C=10, svc__gamma=0.001 .....................................\n",
      "[CV] . svc__C=1, svc__gamma=1, score=0.8629726694122052, total=  13.5s\n",
      "[CV] svc__C=10, svc__gamma=0.001 .....................................\n",
      "[CV]  svc__C=10, svc__gamma=0.0005, score=0.8769173213617658, total=  13.9s\n",
      "[CV] svc__C=10, svc__gamma=0.001 .....................................\n",
      "[CV]  svc__C=10, svc__gamma=0.0005, score=0.7942386831275721, total=  16.6s\n",
      "[CV] svc__C=10, svc__gamma=0.001 .....................................\n",
      "[CV]  svc__C=10, svc__gamma=0.001, score=0.8903853348297793, total=  13.4s\n",
      "[CV] svc__C=10, svc__gamma=0.01 ......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.0005, score=0.8244760479041916, total=  16.8s\n",
      "[CV] svc__C=10, svc__gamma=0.01 ......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.0005, score=0.8181137724550899, total=  16.9s\n",
      "[CV] svc__C=10, svc__gamma=0.01 ......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.0005, score=0.8172968925496069, total=  17.2s\n",
      "[CV] svc__C=10, svc__gamma=0.01 ......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.001, score=0.8065843621399177, total=  16.2s\n",
      "[CV] svc__C=10, svc__gamma=0.01 ......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.001, score=0.8327095808383234, total=  15.4s\n",
      "[CV] svc__C=10, svc__gamma=0.1 .......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.01, score=0.9176954732510288, total=  11.3s\n",
      "[CV] svc__C=10, svc__gamma=0.1 .......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.01, score=0.8555929667040778, total=  14.2s\n",
      "[CV] svc__C=10, svc__gamma=0.1 .......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.001, score=0.8308383233532934, total=  16.8s\n",
      "[CV] svc__C=10, svc__gamma=0.1 .......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.01, score=0.8768712574850299, total=  13.8s\n",
      "[CV] svc__C=10, svc__gamma=0.1 .......................................\n",
      "[CV]  svc__C=10, svc__gamma=0.01, score=0.8566616766467066, total=  14.3s\n",
      "[CV]  svc__C=10, svc__gamma=0.01, score=0.8573567952077873, total=  14.0s\n",
      "[CV] svc__C=10, svc__gamma=1 .........................................\n",
      "[CV] svc__C=10, svc__gamma=1 .........................................\n",
      "[CV]  svc__C=10, svc__gamma=0.001, score=0.8296518157993261, total=  17.1s\n",
      "[CV] svc__C=10, svc__gamma=1 .........................................\n",
      "[CV]  svc__C=10, svc__gamma=0.1, score=0.9259259259259259, total=   9.5s\n",
      "[CV] svc__C=10, svc__gamma=1 .........................................\n",
      "[CV]  svc__C=10, svc__gamma=0.1, score=0.8600823045267489, total=  13.0s\n",
      "[CV] svc__C=10, svc__gamma=1 .........................................\n",
      "[CV]  svc__C=10, svc__gamma=1, score=0.9266741488963711, total=  10.9s\n",
      "[CV] svc__C=100, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=10, svc__gamma=0.1, score=0.8697604790419161, total=  12.8s\n",
      "[CV] svc__C=100, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=10, svc__gamma=0.1, score=0.8645209580838323, total=  12.7s\n",
      "[CV] svc__C=100, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=10, svc__gamma=0.1, score=0.871209284912018, total=  12.7s\n",
      "[CV] svc__C=100, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=10, svc__gamma=1, score=0.8873502994011976, total=  13.6s\n",
      "[CV] svc__C=100, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=10, svc__gamma=1, score=0.8829031051253273, total=  14.2s\n",
      "[CV] svc__C=100, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=10, svc__gamma=1, score=0.8888056907525271, total=  13.8s\n",
      "[CV] svc__C=100, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=10, svc__gamma=1, score=0.8828592814371258, total=  14.0s\n",
      "[CV] svc__C=100, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.0005, score=0.9075944631500187, total=  12.4s\n",
      "[CV] svc__C=100, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.0005, score=0.843937125748503, total=  15.5s\n",
      "[CV] svc__C=100, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.0005, score=0.8338945005611672, total=  16.0s\n",
      "[CV] svc__C=100, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.0005, score=0.8476796407185628, total=  15.9s\n",
      "[CV] svc__C=100, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.001, score=0.8537224092779648, total=  14.5s\n",
      "[CV] svc__C=100, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.001, score=0.9173213617658063, total=  10.9s\n",
      "[CV] svc__C=100, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.0005, score=0.8457506551853239, total=  15.7s\n",
      "[CV] svc__C=100, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=100, svc__gamma=0.001, score=0.8648952095808383, total=  13.7s\n",
      "[CV] svc__C=100, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=100, svc__gamma=0.01, score=0.9334081556303778, total=   8.0s\n",
      "[CV] svc__C=100, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=100, svc__gamma=0.001, score=0.8615269461077845, total=  14.0s\n",
      "[CV] svc__C=100, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=100, svc__gamma=0.01, score=0.8806584362139918, total=  11.5s\n",
      "[CV] svc__C=100, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=100, svc__gamma=0.01, score=0.8948353293413174, total=  11.6s\n",
      "[CV] svc__C=100, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=100, svc__gamma=0.001, score=0.8539872706851367, total=  13.9s\n",
      "[CV] svc__C=100, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=100, svc__gamma=0.01, score=0.8854790419161677, total=  12.0s\n",
      "[CV] svc__C=100, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=100, svc__gamma=0.01, score=0.8831898165481094, total=  11.7s\n",
      "[CV] svc__C=100, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=100, svc__gamma=0.1, score=0.9375233819678264, total=   7.7s\n",
      "[CV] svc__C=100, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=100, svc__gamma=0.1, score=0.8881406659184437, total=  12.0s\n",
      "[CV] svc__C=100, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=100, svc__gamma=0.1, score=0.9004491017964071, total=  11.8s\n",
      "[CV] svc__C=500, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=100, svc__gamma=1, score=0.9438832772166106, total=   9.9s\n",
      "[CV] svc__C=500, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=100, svc__gamma=0.1, score=0.8948353293413174, total=  12.0s\n",
      "[CV] svc__C=500, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=100, svc__gamma=0.1, score=0.8955447397978286, total=  12.0s\n",
      "[CV] svc__C=500, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=100, svc__gamma=1, score=0.9023569023569024, total=  12.9s\n",
      "[CV] svc__C=500, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=100, svc__gamma=1, score=0.906062874251497, total=  13.7s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] svc__C=500, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=100, svc__gamma=1, score=0.905314371257485, total=  13.1s\n",
      "[CV] svc__C=500, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=100, svc__gamma=1, score=0.8989142643204793, total=  12.8s\n",
      "[CV] svc__C=500, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.0005, score=0.9225589225589226, total=  10.0s\n",
      "[CV] svc__C=500, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.0005, score=0.8649457538346427, total=  14.4s\n",
      "[CV] svc__C=500, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.0005, score=0.874625748502994, total=  14.2s\n",
      "[CV] svc__C=500, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.001, score=0.9270482603815937, total=  10.1s\n",
      "[CV] svc__C=500, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.0005, score=0.8708832335329342, total=  14.4s\n",
      "[CV] svc__C=500, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.001, score=0.8750467639356528, total=  13.5s\n",
      "[CV] svc__C=500, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.0005, score=0.8655934107076001, total=  15.0s\n",
      "[CV] svc__C=500, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=500, svc__gamma=0.001, score=0.8866017964071856, total=  13.5s\n",
      "[CV] svc__C=500, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=500, svc__gamma=0.01, score=0.9393939393939394, total=   7.3s\n",
      "[CV] svc__C=500, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=500, svc__gamma=0.001, score=0.8806137724550899, total=  13.6s\n",
      "[CV] svc__C=500, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=500, svc__gamma=0.01, score=0.9019827908716798, total=  11.6s\n",
      "[CV] svc__C=500, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=500, svc__gamma=0.001, score=0.8745788094346687, total=  13.2s\n",
      "[CV] svc__C=500, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=500, svc__gamma=0.01, score=0.9191616766467066, total=  11.8s\n",
      "[CV] svc__C=500, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=500, svc__gamma=0.01, score=0.9041557469112692, total=  11.4s\n",
      "[CV] svc__C=500, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=500, svc__gamma=0.1, score=0.9476243920688365, total=   8.3s\n",
      "[CV] svc__C=500, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=500, svc__gamma=0.01, score=0.9109281437125748, total=  11.8s\n",
      "[CV] svc__C=500, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=500, svc__gamma=0.1, score=0.9094650205761317, total=  13.0s\n",
      "[CV] svc__C=500, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=500, svc__gamma=0.1, score=0.9172904191616766, total=  12.3s\n",
      "[CV] svc__C=750, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=500, svc__gamma=0.1, score=0.9078996630475478, total=  12.1s\n",
      "[CV]  svc__C=500, svc__gamma=0.1, score=0.9131736526946108, total=  12.9s\n",
      "[CV] svc__C=750, svc__gamma=0.0005 ...................................\n",
      "[CV] svc__C=750, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=500, svc__gamma=1, score=0.9386457164234941, total=  10.9s\n",
      "[CV] svc__C=750, svc__gamma=0.0005 ...................................\n",
      "[CV]  svc__C=500, svc__gamma=1, score=0.9143284698840254, total=  16.2s\n",
      "[CV]  svc__C=500, svc__gamma=1, score=0.9161676646706587, total=  15.0s\n",
      "[CV] svc__C=750, svc__gamma=0.0005 ...................................\n",
      "[CV] svc__C=750, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=500, svc__gamma=1, score=0.9210329341317365, total=  15.9s\n",
      "[CV] svc__C=750, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=500, svc__gamma=1, score=0.9071508798202921, total=  14.8s\n",
      "[CV] svc__C=750, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.0005, score=0.9210624766180322, total=   9.6s\n",
      "[CV] svc__C=750, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.0005, score=0.8720538720538721, total=  12.7s\n",
      "[CV] svc__C=750, svc__gamma=0.001 ....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.0005, score=0.8802395209580839, total=  13.4s\n",
      "[CV] svc__C=750, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.001, score=0.9285447063224841, total=   9.2s\n",
      "[CV] svc__C=750, svc__gamma=0.01 .....................................\n",
      "[CV] ....... svc__C=750, svc__gamma=0.0005, score=0.875, total=  13.3s\n",
      "[CV] svc__C=750, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.001, score=0.8787878787878788, total=  12.4s\n",
      "[CV] svc__C=750, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.0005, score=0.8715836765256458, total=  13.9s\n",
      "[CV] svc__C=750, svc__gamma=0.01 .....................................\n",
      "[CV]  svc__C=750, svc__gamma=0.001, score=0.8914670658682635, total=  12.7s\n",
      "[CV] svc__C=750, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=750, svc__gamma=0.01, score=0.9438832772166106, total=   7.4s\n",
      "[CV] svc__C=750, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=750, svc__gamma=0.001, score=0.8809434668663422, total=  13.1s\n",
      "[CV] svc__C=750, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=750, svc__gamma=0.001, score=0.8851047904191617, total=  13.2s\n",
      "[CV] svc__C=750, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=750, svc__gamma=0.01, score=0.9083426861204639, total=  12.2s\n",
      "[CV] svc__C=750, svc__gamma=0.1 ......................................\n",
      "[CV]  svc__C=750, svc__gamma=0.01, score=0.9202844311377245, total=  12.9s\n",
      "[CV] svc__C=750, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=750, svc__gamma=0.01, score=0.9157934131736527, total=  12.5s\n",
      "[CV] svc__C=750, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=750, svc__gamma=0.1, score=0.9479985035540591, total=   9.0s\n",
      "[CV] svc__C=750, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=750, svc__gamma=0.01, score=0.909771621115687, total=  13.0s\n",
      "[CV] svc__C=750, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=750, svc__gamma=0.1, score=0.9135802469135802, total=  13.4s\n",
      "[CV] svc__C=750, svc__gamma=1 ........................................\n",
      "[CV]  svc__C=750, svc__gamma=0.1, score=0.9217814371257484, total=  12.7s\n",
      "[CV]  svc__C=750, svc__gamma=0.1, score=0.9157934131736527, total=  13.2s\n",
      "[CV]  svc__C=750, svc__gamma=0.1, score=0.9112691875701985, total=  13.0s\n",
      "[CV]  svc__C=750, svc__gamma=1, score=0.9349046015712682, total=  10.8s\n",
      "[CV]  svc__C=750, svc__gamma=1, score=0.9165731387953611, total=  16.0s\n",
      "[CV]  svc__C=750, svc__gamma=1, score=0.9225299401197605, total=  15.4s\n",
      "[CV]  svc__C=750, svc__gamma=1, score=0.9127994011976048, total=  13.0s\n",
      "[CV]  svc__C=750, svc__gamma=1, score=0.9127667540247099, total=  12.9s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done 125 out of 125 | elapsed:  5.3min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('categoricaltransformer', CategoricalTransformer(columns=['favorite', 'interested', 'status', 'property_type', 'sale_type', 'source', 'state'])), ('dummyencoder', DummyEncoder()), ('imputer', Imputer(axis=0, copy=True, missing_values='NaN', strategy='median', verbose=0)), ('standardscaler', ...f',\n",
       "  max_iter=-1, probability=False, random_state=42, shrinking=True,\n",
       "  tol=0.001, verbose=False))]),\n",
       "       fit_params=None, iid=True, n_jobs=-1,\n",
       "       param_grid={'svc__C': [1, 10, 100, 500, 750], 'svc__gamma': [0.0005, 0.001, 0.01, 0.1, 1]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=None, verbose=3)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svmgrid = GridSearchCV(svc_pipe, param_grid, cv=5, n_jobs=-1, verbose=3)#, scoring=roc_auc_scorer)\n",
    "svmgrid.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.90180396517871109"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_test, svmgrid.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 748.8x514.8 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "scores = svmgrid.cv_results_['mean_test_score'].reshape(5,5)\n",
    "sns.heatmap(scores, vmax=1, xticklabels=param_grid['svc__gamma'], yticklabels=param_grid['svc__C'], cmap='hot', annot=True)\n",
    "plt.title('Hyper parameters vs score')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'SVM')"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 748.8x514.8 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_roc(svmgrid, X_test, y_test)\n",
    "plt.title('SVM')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Adaboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "n_estimators = [ 1500, 1570, 2000, 2250, 3000, 3250, 3500]\n",
    "param_grid = {'adaboostclassifier__n_estimators': n_estimators}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "abpipe = make_pipeline(CategoricalTransformer(columns=cat_cols), DummyEncoder(), Imputer(strategy='median'), StandardScaler(), \n",
    "                   AdaBoostClassifier())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "scrolled": true,
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 7 candidates, totalling 35 fits\n",
      "[CV] adaboostclassifier__n_estimators=1500 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1500 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1500 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1500 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1500 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1570 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1570 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=1570 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1500, score=0.9974806405731442, total=  19.8s\n",
      "[CV] adaboostclassifier__n_estimators=1570 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1500, score=0.996398639757171, total=  19.9s\n",
      "[CV] adaboostclassifier__n_estimators=1570 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1500, score=0.997032827451858, total=  20.0s\n",
      "[CV] adaboostclassifier__n_estimators=2000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1500, score=0.9964454477248496, total=  20.2s\n",
      "[CV] adaboostclassifier__n_estimators=2000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1500, score=0.9957317073170731, total=  20.4s\n",
      "[CV] adaboostclassifier__n_estimators=2000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1570, score=0.9970124254669422, total=  20.5s\n",
      "[CV] adaboostclassifier__n_estimators=2000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1570, score=0.9974924932834642, total=  21.0s\n",
      "[CV] adaboostclassifier__n_estimators=2000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1570, score=0.99648032853777, total=  21.3s\n",
      "[CV] adaboostclassifier__n_estimators=2250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=1570, score=0.996334029541373, total=  22.3s\n",
      "[CV]  adaboostclassifier__n_estimators=1570, score=0.9957521203181793, total=  22.4s\n",
      "[CV] adaboostclassifier__n_estimators=2250 ...........................\n",
      "[CV] adaboostclassifier__n_estimators=2250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2000, score=0.9971358245692549, total=  28.1s\n",
      "[CV] adaboostclassifier__n_estimators=2250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2000, score=0.9965517354849749, total=  28.7s\n",
      "[CV] adaboostclassifier__n_estimators=2250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2000, score=0.9976037770636885, total=  28.7s\n",
      "[CV] adaboostclassifier__n_estimators=3000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2000, score=0.9958703181794236, total=  28.8s\n",
      "[CV] adaboostclassifier__n_estimators=3000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2000, score=0.9963452374359503, total=  28.4s\n",
      "[CV] adaboostclassifier__n_estimators=3000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2250, score=0.9965122477722349, total=  32.4s\n",
      "[CV] adaboostclassifier__n_estimators=3000 ...........................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:  1.0min\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  adaboostclassifier__n_estimators=2250, score=0.9971371408263462, total=  32.3s\n",
      "[CV] adaboostclassifier__n_estimators=3000 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2250, score=0.9975270636885634, total=  32.6s\n",
      "[CV] adaboostclassifier__n_estimators=3250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2250, score=0.9959233261339093, total=  33.4s\n",
      "[CV] adaboostclassifier__n_estimators=3250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=2250, score=0.9962773307805706, total=  33.3s\n",
      "[CV] adaboostclassifier__n_estimators=3250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3000, score=0.9971509615258052, total=  43.3s\n",
      "[CV] adaboostclassifier__n_estimators=3250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3000, score=0.9966264330749082, total=  44.3s\n",
      "[CV] adaboostclassifier__n_estimators=3250 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3000, score=0.9975313438339567, total=  44.7s\n",
      "[CV] adaboostclassifier__n_estimators=3500 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3000, score=0.9959674445556551, total=  44.5s\n",
      "[CV] adaboostclassifier__n_estimators=3500 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3000, score=0.9962789790003613, total=  44.2s\n",
      "[CV] adaboostclassifier__n_estimators=3500 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3250, score=0.996706066628934, total=  47.6s\n",
      "[CV] adaboostclassifier__n_estimators=3500 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3250, score=0.9972101930949153, total=  47.5s\n",
      "[CV] adaboostclassifier__n_estimators=3500 ...........................\n",
      "[CV]  adaboostclassifier__n_estimators=3250, score=0.9975178449138703, total=  48.4s\n",
      "[CV]  adaboostclassifier__n_estimators=3250, score=0.9962516185518345, total=  47.4s\n",
      "[CV]  adaboostclassifier__n_estimators=3250, score=0.9959878575567613, total=  48.3s\n",
      "[CV]  adaboostclassifier__n_estimators=3500, score=0.9966935621865662, total=  50.7s\n",
      "[CV]  adaboostclassifier__n_estimators=3500, score=0.997175312281995, total=  49.8s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  32 out of  35 | elapsed:  2.8min remaining:   16.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  adaboostclassifier__n_estimators=3500, score=0.9975040167518305, total=  41.5s\n",
      "[CV]  adaboostclassifier__n_estimators=3500, score=0.9959628351683085, total=  39.6s\n",
      "[CV]  adaboostclassifier__n_estimators=3500, score=0.9962746936289053, total=  36.2s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  35 out of  35 | elapsed:  3.2min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=Pipeline(memory=None,\n",
       "     steps=[('categoricaltransformer', CategoricalTransformer(columns=['favorite', 'interested', 'status', 'property_type', 'sale_type', 'source', 'state'])), ('dummyencoder', DummyEncoder()), ('imputer', Imputer(axis=0, copy=True, missing_values='NaN', strategy='median', verbose=0)), ('standardscaler', ...m='SAMME.R', base_estimator=None,\n",
       "          learning_rate=1.0, n_estimators=50, random_state=None))]),\n",
       "       fit_params=None, iid=True, n_jobs=-1,\n",
       "       param_grid={'adaboostclassifier__n_estimators': [1500, 1570, 2000, 2250, 3000, 3250, 3500]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=make_scorer(roc_auc_score, needs_threshold=True), verbose=3)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abgrid = GridSearchCV(abpipe, param_grid, cv=5, n_jobs=-1, verbose=3, scoring=roc_auc_scorer)\n",
    "abgrid.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f22bbfd0ef0>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 748.8x514.8 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pd.DataFrame({'score':abgrid.cv_results_['mean_test_score']}, index=abgrid.cv_results_['param_adaboostclassifier__n_estimators']).plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.96757739512200769"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_test, abgrid.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Pipeline(memory=None,\n",
       "      steps=[('categoricaltransformer', CategoricalTransformer(columns=['favorite', 'interested', 'status', 'property_type', 'sale_type', 'source', 'state'])), ('dummyencoder', DummyEncoder()), ('imputer', Imputer(axis=0, copy=True, missing_values='NaN', strategy='median', verbose=0)), ('standardscaler', ...'SAMME.R', base_estimator=None,\n",
       "           learning_rate=1.0, n_estimators=3250, random_state=None))]),\n",
       " {'adaboostclassifier__n_estimators': 3250})"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abgrid.best_estimator_, abgrid.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Thank you!"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  },
  "livereveal": {
   "theme": "white",
   "transition": "None"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}