jannik.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Getting to know the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "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>diagnosis</th>\n",
       "      <th>radius_mean</th>\n",
       "      <th>texture_mean</th>\n",
       "      <th>perimeter_mean</th>\n",
       "      <th>area_mean</th>\n",
       "      <th>smoothness_mean</th>\n",
       "      <th>compactness_mean</th>\n",
       "      <th>concavity_mean</th>\n",
       "      <th>concave points_mean</th>\n",
       "      <th>symmetry_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>radius_worst</th>\n",
       "      <th>texture_worst</th>\n",
       "      <th>perimeter_worst</th>\n",
       "      <th>area_worst</th>\n",
       "      <th>smoothness_worst</th>\n",
       "      <th>compactness_worst</th>\n",
       "      <th>concavity_worst</th>\n",
       "      <th>concave points_worst</th>\n",
       "      <th>symmetry_worst</th>\n",
       "      <th>fractal_dimension_worst</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>M</td>\n",
       "      <td>17.99</td>\n",
       "      <td>10.38</td>\n",
       "      <td>122.80</td>\n",
       "      <td>1001.0</td>\n",
       "      <td>0.11840</td>\n",
       "      <td>0.27760</td>\n",
       "      <td>0.3001</td>\n",
       "      <td>0.14710</td>\n",
       "      <td>0.2419</td>\n",
       "      <td>...</td>\n",
       "      <td>25.38</td>\n",
       "      <td>17.33</td>\n",
       "      <td>184.60</td>\n",
       "      <td>2019.0</td>\n",
       "      <td>0.1622</td>\n",
       "      <td>0.6656</td>\n",
       "      <td>0.7119</td>\n",
       "      <td>0.2654</td>\n",
       "      <td>0.4601</td>\n",
       "      <td>0.11890</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>M</td>\n",
       "      <td>20.57</td>\n",
       "      <td>17.77</td>\n",
       "      <td>132.90</td>\n",
       "      <td>1326.0</td>\n",
       "      <td>0.08474</td>\n",
       "      <td>0.07864</td>\n",
       "      <td>0.0869</td>\n",
       "      <td>0.07017</td>\n",
       "      <td>0.1812</td>\n",
       "      <td>...</td>\n",
       "      <td>24.99</td>\n",
       "      <td>23.41</td>\n",
       "      <td>158.80</td>\n",
       "      <td>1956.0</td>\n",
       "      <td>0.1238</td>\n",
       "      <td>0.1866</td>\n",
       "      <td>0.2416</td>\n",
       "      <td>0.1860</td>\n",
       "      <td>0.2750</td>\n",
       "      <td>0.08902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>M</td>\n",
       "      <td>19.69</td>\n",
       "      <td>21.25</td>\n",
       "      <td>130.00</td>\n",
       "      <td>1203.0</td>\n",
       "      <td>0.10960</td>\n",
       "      <td>0.15990</td>\n",
       "      <td>0.1974</td>\n",
       "      <td>0.12790</td>\n",
       "      <td>0.2069</td>\n",
       "      <td>...</td>\n",
       "      <td>23.57</td>\n",
       "      <td>25.53</td>\n",
       "      <td>152.50</td>\n",
       "      <td>1709.0</td>\n",
       "      <td>0.1444</td>\n",
       "      <td>0.4245</td>\n",
       "      <td>0.4504</td>\n",
       "      <td>0.2430</td>\n",
       "      <td>0.3613</td>\n",
       "      <td>0.08758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>M</td>\n",
       "      <td>11.42</td>\n",
       "      <td>20.38</td>\n",
       "      <td>77.58</td>\n",
       "      <td>386.1</td>\n",
       "      <td>0.14250</td>\n",
       "      <td>0.28390</td>\n",
       "      <td>0.2414</td>\n",
       "      <td>0.10520</td>\n",
       "      <td>0.2597</td>\n",
       "      <td>...</td>\n",
       "      <td>14.91</td>\n",
       "      <td>26.50</td>\n",
       "      <td>98.87</td>\n",
       "      <td>567.7</td>\n",
       "      <td>0.2098</td>\n",
       "      <td>0.8663</td>\n",
       "      <td>0.6869</td>\n",
       "      <td>0.2575</td>\n",
       "      <td>0.6638</td>\n",
       "      <td>0.17300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>M</td>\n",
       "      <td>20.29</td>\n",
       "      <td>14.34</td>\n",
       "      <td>135.10</td>\n",
       "      <td>1297.0</td>\n",
       "      <td>0.10030</td>\n",
       "      <td>0.13280</td>\n",
       "      <td>0.1980</td>\n",
       "      <td>0.10430</td>\n",
       "      <td>0.1809</td>\n",
       "      <td>...</td>\n",
       "      <td>22.54</td>\n",
       "      <td>16.67</td>\n",
       "      <td>152.20</td>\n",
       "      <td>1575.0</td>\n",
       "      <td>0.1374</td>\n",
       "      <td>0.2050</td>\n",
       "      <td>0.4000</td>\n",
       "      <td>0.1625</td>\n",
       "      <td>0.2364</td>\n",
       "      <td>0.07678</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  diagnosis  radius_mean  texture_mean  perimeter_mean  area_mean  \\\n",
       "0         M        17.99         10.38          122.80     1001.0   \n",
       "1         M        20.57         17.77          132.90     1326.0   \n",
       "2         M        19.69         21.25          130.00     1203.0   \n",
       "3         M        11.42         20.38           77.58      386.1   \n",
       "4         M        20.29         14.34          135.10     1297.0   \n",
       "\n",
       "   smoothness_mean  compactness_mean  concavity_mean  concave points_mean  \\\n",
       "0          0.11840           0.27760          0.3001              0.14710   \n",
       "1          0.08474           0.07864          0.0869              0.07017   \n",
       "2          0.10960           0.15990          0.1974              0.12790   \n",
       "3          0.14250           0.28390          0.2414              0.10520   \n",
       "4          0.10030           0.13280          0.1980              0.10430   \n",
       "\n",
       "   symmetry_mean           ...             radius_worst  texture_worst  \\\n",
       "0         0.2419           ...                    25.38          17.33   \n",
       "1         0.1812           ...                    24.99          23.41   \n",
       "2         0.2069           ...                    23.57          25.53   \n",
       "3         0.2597           ...                    14.91          26.50   \n",
       "4         0.1809           ...                    22.54          16.67   \n",
       "\n",
       "   perimeter_worst  area_worst  smoothness_worst  compactness_worst  \\\n",
       "0           184.60      2019.0            0.1622             0.6656   \n",
       "1           158.80      1956.0            0.1238             0.1866   \n",
       "2           152.50      1709.0            0.1444             0.4245   \n",
       "3            98.87       567.7            0.2098             0.8663   \n",
       "4           152.20      1575.0            0.1374             0.2050   \n",
       "\n",
       "   concavity_worst  concave points_worst  symmetry_worst  \\\n",
       "0           0.7119                0.2654          0.4601   \n",
       "1           0.2416                0.1860          0.2750   \n",
       "2           0.4504                0.2430          0.3613   \n",
       "3           0.6869                0.2575          0.6638   \n",
       "4           0.4000                0.1625          0.2364   \n",
       "\n",
       "   fractal_dimension_worst  \n",
       "0                  0.11890  \n",
       "1                  0.08902  \n",
       "2                  0.08758  \n",
       "3                  0.17300  \n",
       "4                  0.07678  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load data set\n",
    "df = pd.read_csv(\"Wisconsin.csv\")\n",
    "\n",
    "# drop patient id from dataset\n",
    "df.drop('id', axis=1, inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(569, 31)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "B    357\n",
       "M    212\n",
       "Name: diagnosis, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels = ['0', '1']\n",
    "balance = df.diagnosis.value_counts()\n",
    "balance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data set is slightly imbalanced, majority of cells are benign."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.62741652021089633"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "balance[0]/balance.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 1.  1.  1.  1.  1.  1.  1.  1.]\n",
      "[[  1.80e+01   1.04e+01   1.23e+02   1.00e+03   1.18e-01]\n",
      " [  2.06e+01   1.78e+01   1.33e+02   1.33e+03   8.47e-02]\n",
      " [  1.97e+01   2.12e+01   1.30e+02   1.20e+03   1.10e-01]\n",
      " [  1.14e+01   2.04e+01   7.76e+01   3.86e+02   1.42e-01]\n",
      " [  2.03e+01   1.43e+01   1.35e+02   1.30e+03   1.00e-01]\n",
      " [  1.24e+01   1.57e+01   8.26e+01   4.77e+02   1.28e-01]\n",
      " [  1.82e+01   2.00e+01   1.20e+02   1.04e+03   9.46e-02]\n",
      " [  1.37e+01   2.08e+01   9.02e+01   5.78e+02   1.19e-01]]\n",
      "(569,) (569, 30)\n"
     ]
    }
   ],
   "source": [
    "# split into labels Y and instances X\n",
    "Y = df.values[:, 0]\n",
    "# 1 is positive => malign => patient has\n",
    "Y = np.array(list(1. if i=='M' else 0. for i in Y)).astype('float')\n",
    "X = df.values[:, 1:].astype('float64')\n",
    "%precision 2\n",
    "print(Y[0:8])\n",
    "print(X[0:8, :5])\n",
    "print(Y.shape, X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(212, 30)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((424, 30), (424,))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# balanca data, might be useful later\n",
    "X_false = X[Y==0]\n",
    "X_true = X[Y==1]\n",
    "print(X_true.shape)\n",
    "X_balanced = np.concatenate([X_false[0:212], X_true])\n",
    "Y_balanced = np.array(212*[0]+212*[1])\n",
    "\n",
    "X_balanced.shape, Y_balanced.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Scatter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# get random pairs of features\n",
    "lim = X.shape[1]\n",
    "a = np.stack(np.meshgrid(np.arange(0,lim), np.arange(0,lim)), -1).reshape(-1, 2)\n",
    "a = np.array([i for i in a if i[0] != i[1]])\n",
    "np.random.seed(636)\n",
    "inds = np.random.choice(range(0, len(a)), size=16)\n",
    "a = a[inds]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArMAAAGRCAYAAAB2azu7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIYQQLUvCrBBCCCGEaFn/BxXOybHBuAaMAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11675e470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,7))\n",
    "plt.rcParams.update({'font.size': 12})\n",
    "\n",
    "for i in range(0,15):\n",
    "    fig = plt.subplot(3,5,i+1)\n",
    "    plt.xticks([])\n",
    "    plt.yticks([])\n",
    "    \n",
    "    for j, label in enumerate(labels):\n",
    "        plt.scatter(X[Y==j, a[i,0]], X[Y == j, a[i,1]], alpha=.2,\n",
    "                    label=label)\n",
    "    plt.text(0.25,0.9, \"$({}, {})$\".format(a[i,0],a[i,1]),\n",
    "         horizontalalignment='center',\n",
    "         verticalalignment='center',\n",
    "         transform = plt.gca().transAxes)\n",
    "    if i == 14:\n",
    "        plt.legend(labels, loc='lower right')\n",
    "    \n",
    "plt.subplots_adjust(wspace=0, hspace=0)    \n",
    "plt.savefig('figs/random_scatter.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# standardise X\n",
    "scaler = StandardScaler()\n",
    "X_scaled = scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARQAAAEUCAYAAADqcMl5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsvXl8ZGd55/t9z1L7oqW0tbpb6nYv\nlt12G7ubxcZmHwzGxDDJJc4Ehgu5zupkIIHsJJlLwjI3IZ9rxsl4hrlJTFgSiMGOcYMxBoMxxhu2\n5Vbvu/a19u2c894/3iqpVC21St1V2vr8Ph99Sjqn6pxX3Xp/51l+z/MIKSUuXLhwUQ9oq70AFy5c\nbBy4hOLChYu6wSUUFy5c1A0uobhw4aJucAnFhQsXdYNLKC5cuKgbXEJx4cJF3eASigsXLuoGl1Bc\nuHBRNxirvYB6IBaLyd7e3tVehgsXGxbPPffchJSyban3bQhC6e3t5dlnn13tZbhwsWEhhDhdy/tc\nl8eFCxd1g0soLly4qBtcQnHhwkXd4BKKCxcu6oYNEZR14WI14TgO586dI51Or/ZSLhqmadLe3k4k\nErmk67iE4mJdYGA4zoH+UQZnsnQ3+bl1Twd9XdHVXhYAExMTCCHYvXs3mrb+jH4pJdlslsHBQYBL\nIpX199u7uOwwMBznvidOEs8W6Yr6iGeL3PfESQaG46u9NABmZmbo6OhYl2QCIIQgEAjQ3d3N2NjY\nJV1rff4LuLiscKB/lKjfJOo30YSY/f5A/+hqLw0A27YxTXO1l3HJ8Pv9FIvFS7qGSygu1jwGZ7KE\nffO987DPYHAmu0orOh9CiNVewiWjHr+DSygu1jy6m/wkc9a8Y8mcRXeTf5VW5GIxuITiYs3j1j0d\nxLNF4tkijpSz39+6p2O1l7YuMDU1xXve8x6CwSA9PT186Utfati9XEJxsebR1xXlrlu2EfWbDMdz\nRP0md92ybc1kedY6fvM3fxOPx8Po6Cj//M//zK//+q/zyiuvNORebtrYxbpAX1fUJZCLQDqd5utf\n/zr9/f2EQiFe//rX8+53v5v777+fT3/603W/n0soLlysIhqtrzly5Ai6rrNr167ZY3v37uUHP/hB\n3e5RCdflceFilbAS+ppUKkU0Op+gotEoyWSybveohEsoLlysElZCXxMKhUgkEvOOJRIJwuFw3e5R\nCZdQXLhYJayEvmbXrl1YlsXRo0dnj7344otcffXVdbtHJVxCceFilbAS+ppgMMh73/tePvGJT5BO\np3nyySf55je/yfvf//663aMSLqG4cLFKWCl9zb333ks2m6W9vZ0777yTv/u7v2uYheJmeVy4WCWU\n9TWVWZ737d9c9/R4S0sL3/jGN+p6zcXgEooLF6uIjaavcV0eFy5c1A0uobhw4aJucAnFhQsXdYNL\nKC5cuKgbXEJx4cJF3eASigsXLuoGl1BcuHBRN7g6FBfrAyP9MPAQxM9CdAv03Q6de1Z7VS6q4Foo\nLtY+Rvrhx/dAdgYi3er1x/eo4y6WxOc//3n27duH1+vlgx/8YEPvteKEIoT4LSHEs0KIvBDiH6rO\nvUUIcUgIkRFCPC6E6Fnp9blYgxh4CHxN4G8CoalXX5M67mJJbNq0iT/5kz/hQx/6UMPvtRoWyhDw\nSeB/Vx4UQsSAfwP+FGgBngW+uuKrc7H2ED8Lvqppdr6IOu5iSbz3ve/ljjvuoLW1teH3WvEYipTy\n3wCEEPuAzRWn3gu8IqX819L5PwcmhBBXSikPrfQ6XawhRLcoN8ffNHcsl1DH1zs2WGxoLcVQrgZe\nLP8gpUwDx0vHXVzO6LsdcjOKVKSjXnMz6vh6xgaMDa0lQgkB1c0048CCveqEEHeVYjHPjo+PN3xx\nLlYRnXvgxruVhZIYVK833r2un+TAhowNraW0cQqoHvseARbspiulvA+4D2Dfvn2ysUtzsero3LP+\nCaQa8bPKMqnEOo8NrSUL5RVgb/kHIUQQuKJ03IWLjYfoFhULqkQDYkOWZZHL5bBtG9u2yeVyWJa1\n9AcvAquRNjaEED5AB3QhhE8IYQAPAHuEEP+xdP4TwEtuQNbFhsUKxYY++clP4vf7+fSnP80Xv/hF\n/H4/n/zkJ+t6jzKElCvrLZSyN39WdfgvpJR/LoR4K/B5oAd4GviglPLUUtfct2+ffPbZZ+u9VBc1\notHDqtY6BgYG6Ovru7gPr7Esz2K/ixDiOSnlvqU+vxpp4z8H/nyRc98FrlzJ9bi4NJSHVUX95rxh\nVe7s4RqxwWJDaymG4mIdYiWGVblYP3AJxcUlYSWGVblYP3AJxcUlYSWGVblYP3AJxcUlYaWGVa11\nrHRyoxGox+/gEoqLS0J5WFXUbzIczxH1m5ddQFbXdYrF4mov45KRzWYxTfOSrrGWlLIu1ik22rCq\n5aKpqYnR0VG6u7vRtPX3jJZSks1mGRwcpKPj0ixLl1BcuLhExGIxzp07x+HDh1d7KRcN0zTp6Ogg\nEqmuflkeXELZYLjcRWarAU3T2Lp162ovY01g/dlnLhZFWWQWzxbnicwGhquLuF24aAxcQtlAcEVm\nLlYbrsuzgTA4k6Ur6pt3bF2KzNZYfYuL2uFaKBsIG0JktgG7mF1OcAllA2FDiMw2YBezywmuy7OB\nUBaZHegf5eBwnHjWIuIzZmMoaz7bM9IPhx4GKcEfhdhuCLWv+y5mlxNcC2WDoa8ryq17Ogh5Ta7q\nitDXFVkf2Z6yq6N7wPBAMQdnn4bU2MbpcH8ZwCWUDYh1me0puzpde8EuqGO6F4Zf3Bgd7i8TuC7P\nGsfFCNXWZban3LBZaLD51TBxGHJxQGyMDveXCVwLZQ3jYoVq6zLbU9mwOdQOvTdDz+vhyttcMllH\ncAllDaPadSlYNifGU/zuv7zE5x49siixrMtsz0Yd5nWZwSWUNYzKbmjjyRzPn5kBKXGkc0FrZV22\nFNiow7wuM1wwhiKECAGfBW5Czcf5Uynl8YrzL0spr2nsEi9fdDf5iWeLRP0mx8bTeA3F/1G/h6hf\n9a040D96HlGs2wLBDdaw+XLEUhbKXwO9qC71Y8BPhRBvrDjf24hFuVCodF2S2SJSSvKWw472ILBw\noNUtEHSxmliKUG4H7pRSPiCl/C/ALwL/IoR4Z+OX5qLSdUGAEIIbepqIhVQGZ6FA67pMGbvYMFgq\nbewH8uUfpJSPCiHeDXxDCHE3sP4baa5xlLuh3bqng/ueOImp6zhSksxZxLNF3rd/87z3r8uUsYsN\ng6UslH7glsoDUsqfAO8A7gECDVqXiyrUGmhdlyljFxsGS1konwJaqg9KKV8ojQ39aENW5WJB1NK7\ntWzJgLJMTk+kOTKaYkurn889emQ2dbwug7Yu1jxWfLZxI+DONp6PcpbnlaE456az7O4IsbU1SDJn\ncXoyjSYEW1oChH3GrOu05tPKLlYVa3a2sYvGo2zJfO7RI2xuDsymmKN+k6m0qpPZ0x2dPQYLp59d\nuFguXGHbBsZCY0ILlkPesucdc4O2LuoFl1A2MBYK0HoMDa+hzzvmBm1d1Auuy7OBUR2gTeYsWoIe\nNCGIZ4vzYijV6edF4fZ7dXEB1BSUFUJ4gA8C1wGhynNSyg80ZGXLgBuUXRwLyfDhIrM85SZIvibV\nRS2XUAV8bs3Nhke9g7L/COwFHgJcyeU6wmKp5osKwFb2e4W514GHXEJxAdROKLcC26SUM41cjIs1\njnITpEq4/V5dVKBWQjkDeBu5kMsV66oyOLpF9SkpWybg9nt1MQ+1Znn+CfimEOJOIcSbK78aubiN\njnVXGew2QXKxBGq1UH6r9PpXVcclsL1+y7m8UFkZDOtAZFZuglSZ5bn+/W78xMUsaiIUKeW2Ri/k\ncsS6rAx2myC5uABq1qEIIQzgRqAbOAc8JaW0LvwpFxdCZUe2Mi5bkZmrb9kQqCmGIoS4EhgAvgT8\nNvBl4JAQoq/eCxJCfF8IkRNCpEpfh+t9j7WCddlMuhFw5xlvGNQalL0XuA/YIqV8nZRyM/D3peON\nwG9JKUOlr90NuseagN/UePrkJI8NjFG07Muz6tedZ7xhUKvLcx3wNjlfVvu3wB/Xf0mXB8oZnqjf\n5K19HbMS+Hpev6Z0dD1cjUu9hqtv2TCo1UIZAt5Qdezm0vFG4FNCiAkhxJNVTbFnIYS4SwjxrBDi\n2fHx8QYto3FoZO/XmtPR9XA16nGNyiFfZaxHfctIPzz+KfjGb6jXy9Blq5VQ/gh4UAjxFSHEZ4QQ\nXwEeLB2vN34flYruRrlZDwkhrqh+k5TyPinlPinlvra2tgYso7FYqLVAvTI8NZNVPVyNelxjI+hb\n3DgQUCOhSCkfBK5H9ZgNl15vkFJ+s94LklI+LaVMSinzUsp/BJ4ENlyX/Ub2fq2ZrOJnlWtRieW6\nGvW4xkYY8uXGgYBlpI2llEeATzZwLYveGhCrcN9ZNEIev1BrgWW1EbgAak1Hj+ltnDt6mgnLT9hn\nsqM9SJueW56rUS85/nrXt7hxIOACFooQ4r6K7+8XQvzTQl/1XIwQokkI8XYhhE8IYQgh/hOq6/63\n63mf5aBR8vhGjgutJR09MBznHyavRcvHadWz5AtFDp44y8z0OPTdzsBwnM89eoTf+9cXLzhHeUO4\nK/XARokDXSIuZKGcrPj+WKMXUoKJsoKuBGzgEHCHlHLVtCiNlMfX0sX+Yq971y3b5llV79u/ed69\nDvSPkmm5kpeiH2Dn1OPE8sOMe9t5MPBW9sstsxmoShJdkPBcOb5C3+0qZgLze8Vc//7VXdcKY1FC\nkVJ+quL7v1iJxUgpx4H9K3GvWtFIeXwjK42XIqvy7zUpdjIZ2gmAIyXD8RyTyyXR9e6u1AMusQI1\nxlCEEG8CTkkpTwohOoHPoCyIP5JSjjRygauNRsnjK3UoS1oBDcBiv9e1xjk6jnyHTUzgaB4kYMgC\nCU8nP/bcCOxq+NrWLVxirTkoey/w9tL3f1N6tVBp3XfXe1FrCY0Knq52pXH592rPHOO61BP4M0Ok\nbYNro2mOGTFyeY3tqedAwrnI9Wi5OG9OfZV/esDHS9bm5beOdOt0LgvUqkPpllKeKRUIvh24C/h1\nVLHghkajgqeN1KHUgr6uKHdfneONk1/BzkyTC3Rxvec0kfQZtjWZNGVPkyVAXg8RSZ9ktOhjwvaz\nZfS7ywtOu/qMywq1WigJIUQHsAc4KKVMlRpXm0t8bkPgUoKni8VJFnM5vLrgc48e4ZWhOImcRdRv\ncFVpWHrlGuoRf9k+/jhs2zqX8j3kgBGmOXMSX6DIRMFH1rIJizQhr4Eumog5Y7NiOajBonL70F5W\nqNVCuQd4Bvhn4L+Xjt2EysK4WAQXSjnv6gjykxOT/PtLQzx1fIJTEynOTmUYiuc4OZ7i3FSWRLbI\nmckMpyZS86yBuqWyq0VpvohS/OQS+EPNbAlrXNFs0t7Wji0lYZEl6e2afXtNFlU9hG8u1g1qbbD0\nGSHEA4AtpTxeOjwI/ErDVrbCaETGZbE4yRefOk2m6NAR9nB0LM1AIsnx8TR9HSF2dEY4OJzAa2r4\nTJ1c0WYkkaevKzJrDdz/1GlOjKco2pKQz2BHW3BWWr+sNVeL0gIxOPVDQCi1Z2YKTB90XkOskEXL\nxzna8XOzH68pOH0h4ZsbW9lwqHlyoJTyiJTyuBBCE0JoKG3KK41b2sqh3uK1sijsGz8b5OBQnPFk\nbvZc2Gfwwtk4tu0wmizQFvbS1xWmLezl5eEkuaJFKmfhNdR/jdfQSOSKs9bAwHCcHx2bBCkJeXVm\n0gW+/coo3z80yncOjixvzZWitOQIDD4PjgU4ED8HVh4QcOoJdnGK572v44Teu7zeLYsJ39r63NjK\nBkStaePrUa7OtUBZlCFQsnh9sc+tF9Qz4/LwS4Pc873jpPNFptIFJpI5jo+nec22ZnZ2REjmLCSS\nkUQOr6GsEICIz2A0AQPDSUI+g3zRxmfq5C2HiM+ctQYO9I/SHFDryxRsJtMFBJAp2kQDnuWlniu1\nEy9/DYpptbn9zZCdVqQiWuHqOwjlErx3+iUetHfyUnzzgmK5WVRbHrveAeMD8/UZKxVbca2gFcVy\nBn09BHwIyDRuOauDeonXBobj3PPYcfK2TaZgY2ga2aKNJmyePjmFqWtomsartjTx05PTtATnArJ5\ny2FT1MdoIkfYZzCayOPRNYJeneu2NM2mqr/wo1NctSnMC2fiTKcLGBpIqT6/pzuCqevLI8KyduLQ\nw+DrBbPkwhRSYHghOzlb7NYEfMD/ErzptsWvVzldsGx5HHnk/GK/n9zb+NqXhdby43vWX+HhOkKt\nhNID/HFVg6UNg3qJ17741GlGkzmyBQtdU2Sga4K8ZVOwJMOJPH/x7qsAeP70DImcRcRnkLcc8pZD\nZ8RLpmAT8Bi0BByyRYeCLQl5DX75dT3zskM39DTx7VdGkVJiaIKtLQFiIR+OlBefeq7837XyIPT5\nB5fa8CP98MjHIT0BwRjEdkOoXZ2rtjxWYsaPm2FacdRKKA8A/4FVLNJrJOohXhsYjvPDY5OYuiAD\nSCTJnEXIa+AxdHa2h9jaEpi1HO5+yxXc873jTKYLtAY99LYGODya4pruKNva5sZHl4mu/LnyWqN+\nk+2xIIlSC4RrutX5i1bxbt6vArJCKMtE6FDMQEvFwIPqDV/pTuheSAwpMtG9MHUCBl8EbxC8YWX5\nVLobK1H74lYArzhqDcr6gAeEEN9pZLXxaqEe4rVybKPc1EhKFWRK5YtoQtAZ8c3b6Ldd283n3reX\nd127ia2tQXpjITY3++mJBeddt9r1qlxrNGCChF3tIVpD3ktrcr3/w9BS6mOVS6inuScILTsWriKu\nFqwNvwhTpQRg4hxYBZAW5JOQmVTB3sqg60r0QHErgFcctVooB0tfGxaXWvk7OJOlryvMz87GifoN\nxpOFWWfhqk4fuq6dt9Gr7/m5R4/U5HpVfq463V1zoLQ6ONm5B97yp/Pf09Z3fjC1/JmBh8C2YPRl\ntUlTo+Brgdy0Om/lQdMVGZWtnnLDofI1Gl374lYArzhq1aGsSLXxesBSytftsQBPn8wT9BrYjoMm\nBOmC5K19bUsS1sW4XjURYa3ByQU3+B0LX3P4JZg5BYZfuTSZCciMKcuk9QqYPAqOA4ZHkZFjr4y7\nUUuGyY2fNAzLGfT1NuAXgXYp5e1CiH1AREr5vYatbg1hYDjO/U+d5kfHJmkOmFy1KTyvQrhMBkfH\nUugaOFJg6gav7m3G1DXu/f4JPv/4cQSCV22J8suv6wE4j5yW6mNycYtfIji5lPWy0Pl8HNCU8A0g\n1AFTJ5WfJ3TwRsHOQ9NW9bPpa7y7UWuGyUXDUKsO5W7gd4D/Bfx86XAW+H+5DAoEy8K3E+Mpmv3q\nn+yFM3Fu6GmaVah+5G27uLIzyMMvD4EEjyFoCXgZGE6QtxySeYsrYkEk8NSJKY6OpQj7TLa0BOiK\n+jg5nuIjXxlhS6ufq7qifPj1vXWrH/rg+AmaO3vnv6lsLSxlvSx2HqHcmWJuLogbaFWxkty0er9T\nUG6RLEDL9sa7G25WZ9VRq4XyX4C3SClPCSF+v3TsELAhh3A9/NIg//jUGUYTOToiPmIBk/aoX0nd\nvTpCqBa3x8bSvHpby6yC9YEXhol4TTShwgbpgo10HLKWQ8Rn4PMYpPMW6bzF0EyWqN+kI+JlMmVz\nZCwFAuKZ4iX1Rnn4pUHueew4liNpCZoUijZPTwV4nX2EqDWurARfBIwAWBl44NdA90DX3rnmyjC3\nCRfbpPGz0HE1JIfmrtlyg3J3+m5Xnzv5I4ifVkKZ5BDsv6uxG9vN6qw6aiWUMFD+XynHGk2gUPcV\nrTIefmmQTz9ymKDXoD3kIZEt0j8Y55adrfMUrGVJfDloes93j3JqIo0jJY6UBEwdn0cnkXewHElb\n2Es6bzEcz5VcIoe8ZfP8mRl0TSClSjMPx7MIIeiMePniU6eJhX0cHI4TiR/hFvsprvBM075lB+37\nf+G8zTkwHOee7x0HAS1Bk7zlcGQsxbXeHuS5RyHaTFr4yY2cJVgYZzC8l02Ghd8uKgLwR5Xr0rpz\nbhMutEmtHCRHVZrY3wLdN4Dhm8sCldc1fUp9Xw6IHnkEYjsaRyoroW2pFzaogrdWQnkC+APgLyuO\n/TbweN1XtMr4x6fOEPQaFTJ8jYCp88KZOG/pa+f5MzMASCnx6Fqp/sfD949OABJDA1sKUgUb23EA\nQdRnoGsaE6k8qkRHYOo6XkMR0+nJNJoQCCEIegxyRZsXz86QzNuEvDrh+GHeqT9MUgtxNNzM5Imz\nFMY+w/dafnFes6MD/aMUbYfWoIetxZO8Jv8kzcVROtOjDJrdmLpgemqSoJMnbcbQ7BxjeZtuOYZh\nmMp9Kebg9JPQe7P6BylvUrsAE4cVkWSnFPFseT2MvAwnfwDb3zQ/VrEa7sd6yepsYAVvrTqUu4H3\nCCFOAeHSAPNfAD7aqIWtFkYTOcLe+eVJ7REPqbyFx9C5botyQWayFns2Rbjrlm08dmiCgKkT8BhI\nBIamYWiQtyW6pgr8JlMFMgUbR0rylkPUbxLw6EgpyVsSp2T3tQQ9+Eyd6UyReLbASCLH27VnSBJk\n2g4wkiwykvfwxFkLOfAgZybTs+0NXhmK0xr00Jk9zu2Zr+MpJDhTbMJXmCSYH+PlXCfHmm+haEbI\neFoJyAy6gIItVYrXyp//D9J3u7I0Tv0QClklyXdsRTBCg51vg51vV4rYys1wMW0LLnXy3nqZ77OB\nZ/jUmjYeFkLsB14NbEW5Pz+VUjqNXNxqoCPiI5EtEvXPca0jBdtiqkVAKm/xxt3t89objCZytEc8\njCYK+E2dbNGmaKs4ys07Y3h0nZfOzSClxHIEW5r9XFsipv7BBJqQ6Jog5NWZSucZnMmSzFuzbL9Z\nm2CEGJoAy7YZTzkYmp/d+iR5y+HwaIrdHSESOYuuqI99008y6fgZt3xICTOiiU49h3fmGKm2ZvJG\nmEB+HEMWCdsZitIAu6hIpdSuALtELp17ILIJ0uOKRKQDzb2gGcpiCbUvTBTLdT9qCQ7X4iIspW0Z\n6YdnvgDnnin94+5Xor6VJJ0NHOtZzqAvCTxd+tqw+M+v28qnH1FTO8JenWTeJp23+IN37Oa2a7sX\n/EyZhLqiPqYyBbJFG48hiPhMWkNeXjwbZyZrIYFmv8n2tiAtQS/JnMX2thBXb4owNJ3h5aEkjpT4\nTZ1krogjVaBqULYRFWkSMoBTUuA2a1km9PbZauXheI5owETTNHZ6pzmYCuFIiSYEmdB2vNYALVaC\nw4kc271eQsUJUmYMR/jwkgcH6H09tPeViKBz7he087DjLeppeuqHc5mdsgp1IaJYrvtxIRcJ6uMi\njPTDY/+3UvR6wuof8tQPlTXzlk+sHKmsp1jPMlGTyyOE2CuE+J4QYkoIUSh9FYUQGy4oe9u13fzB\nO3YT8ZuMpQpE/OYsmSw0/GpgOE4sYHJmKsNIPEeT38DUNby6xu6OEN8bGOXoWIps0SJXsJnJFPj+\nkXG+9tw5Dg4neGtfGzftaOXloSSW42BoglSuiOOAIyFnSb5t7yMsU4RIowlJVKRp0bM87b0JKLlU\n6QJXlWT5MrqFsJYlFvSyoz2E2dTFRHAHlq+FSHEMYWU5F34VCaOVgvBimgYE2yAzvvCgrkoJe2y3\nIphcQgnaFhvstVz340IuUr1chIGHlKXljYDHr+qLvBFVf7SS7sYGHo5Wq4XyZeDrqEDsynRRXkXc\ndm33edbIwHCcP/vmK5yaTJMrZXoeGxjBZ+oUbUmT3yCeLXJiwiLsNbi6K8xEush4SnGuI0ttBrIW\npgamJphKFfhvBw7THvHh92jkCpKpdAFbzs1elcAr9la+wG28XX+GLfokeX8XX9dfx7DowSsliZyF\nWZL293VF4bYPcvDrf0lCFpCmD6+VQAqdb2/9XQY929g68hmGnBbCES872oOYJGD8kGqy1HvL+WrS\nSmsjGIPYlUp96i9t8sXUp8uR1l/oqV0vFyF+VsWJKonL8EI+sfC1GpWJ2cAzfGollE7gExu1fUEt\n+MuHD/LCmWk0TWDqGvmiw8BwEp+psaM9zObmAG1hh1TOYldHCJ/H4OmTU1gLRJmKDrOtC0aTeYZK\nBYmF0pt1wayWpWADAo6IXojt4V17Ozk0kiabyjE+liaZs/AYGnfdUiGE69yD7w2/w5nvf5mOzDDZ\nwCaOeK9h+8T3uLMlQ7N/mj1hP7RuKq3Ip4r4hJizCCaOzZesV0rYYzvg5o/UdwNcyEUaeKg+LkJ0\nC4wdUqRSVvhaeVUdXX2tRmdiNugMn+U0WPolVJPqDYel+skODMd55tQMQigykVJScGTJJXHwmTqT\nqRyjiTx5y2EonuXOV28mby/Ov44jOTWVnc3uxLMFiraySAwBEoFA4tEFQsC2WIAbels4NJLmys4g\nD7yQwO/R2dzspyvq49BImoHh+Oy6t+95LfnWPg70j2IPv8w7k1+jo7OD5pZemHTmgpIt25Rk/twz\n0L1PbZ7J4/DiV9TPrdtrk7Bf6tN8qad2PdLBfbfD0M9UDKVcDp5PKhVvtbvhqm4vCrUSyqeBp4QQ\nfwSMVp6QUr657qtaQdQywe9A/yjSkRiGwHEkliOxHYlE1b9NpnKcmcqiawJdgO1IvvbcEF5Do2A5\nLMQrVtWxMpkIlHtkCDB0DcdRR6N+z6w25rFDE7x2e+u8quR4trhop7Zrkk+QM8I43qiKQcR2qBPJ\nIVW8lxxS5FF53BOC1LA6ttRmqtfTfLGndr1chHJFdWWWp/fmhbM8GzgT00jUSihfQw1Pf4ANFkOp\npZ/s4EyWsF8nkVU6Einn5MISOD2ZQdcFmhBYEqJ+g4Lt4DhyQTJZCJXX0wGfIUhbDrYDXgNaS+0i\nVXvIHK/Z1jLv89V9UyqJchMTTDoxnjs9ww09TbSFfMryML1wx71K91HePKkxmDiqgoXpsbmuaxfa\nTBf7NF+OVVMvF6FzD9z+uaXft4EzMY1ErYRyHdAqpdxwWZ3F+sm+MqQyOgeH47wylAApKDpyXrC0\njKIDRUfiMyQ+U6Mj4mMkniVeA5toqIxtGQLQDUGm6CABjwadER8nJjI0Bz2Yuk5HxEcyZ12wb0ol\nUaZ8XUSsBIggx8bStJFQDZEQpHW+AAAgAElEQVTsghKQ6aUUsF2As0+rQj/pqNdzP4XNr1b1Pgtt\nppF+1Y8WCb5obQRU/txaVouuF9XtGkOthPJD4CrgZw1cS8OxUKxkoX6ypyfSnJvOEvToHBlJks5Z\nZAo2mmA25gFgaGBoglzJfynaKu07lSmQK57//oVQHbPd3OQjbzsULMnmZj/pvIWhC6bSeb718ggd\nYR/vub6Ln56c4YVUnrxl4zV0YiEv+6/fxOcePcLgTJbnTk8RMHUcYFB7Fb9kf5OwD/KJNGSUzoae\nm9RGTgwBcq59o79JHQu0KSIZfhFiO2Hr6xQBVTZgOvKIeo+USp+yFAHN/mes8RjFBs7ENBK1EspJ\n4DulYV/VMZRP1H1VDcBisZK39rXx3YFxYK6p0ZHRFLs7Q/QPxhmeyWFLFYAtxzg0wDQ09HJdjmZT\ndNSeshyHiUSOgjOX+l0OBuM5DA3awj7e0tfBkZEEPz01jeNINA12d4b46ckZ4lllLIrSXeLZAk8/\n9SPeJH/Czflh9qfUuI4Wj2RS7+DfuY79+gjXF5+HYERtkMmj6slb3vx2AZAQaIHYLtU0KRtX2Z9d\n71DkUWlRPPFZaL9KVSqffVoVCFYS0EJP87Kb8/K/QKgT2q5UFk1qTKWuUyPqfWuhWK7RmZgNWCBY\nK6EEgIcBD7AuncjFYiVHRtPnNTXa0urH79E5PZXBLpkQlTEOGzCQSCmQKNm8AKxSXY4QKvVb1pMs\nJ9cupUoVj8ZzHB1NMJkpsrlZuTI+U6c3FuL7h8cAeONu1VG+NXWU2MD9XGe9AIFWTuWbeK04hS3h\nhUIfEX+aG62f8rX8e3jt5iyEAiooafiUOK2Yg5EXVYGf7pkfNyjHEcYHFJnYBVU8mEso3YpmwO53\nwJbXKCl+mYAWcl0q3ZxwJ2QTiohadypyQ5SOrzH3pxFY6y7fRaLWWp7/s9ELaTQuNHtnod6u3z88\nNi/4Wo28JdGQsy0NQl6dTNHGdiSaJiiU3KDlCnfKVpCU8NTxSTqjfkxdULAle7ojpXvbs5ZJa+oo\n+4a+iLCOE5chmoC+Yj8JrYm88NLjDHFY7qXJgP+Db9Ocn4GzhxRxhDvU5hdCtSEAFSeYPK4IJJ9Q\nwq/X/TZMHALNnE9ESBgbUN3xQ+0qflImpIXqb1Jjc25O7ErlHkkBg8+q4wJ1fK25P43AWnf5LhKL\nSu+FEL0V329f7GslFlkPdDf5SZZGTpSx2MiJW/d0MJ0pciEZX9nyyFs2Hl2QtyVFW+I4ULTkBYlE\nW8IXkqh9NpO1GE/lyRZtbuhpIhZShOg1dDylUaV7R75GU/Y03XKIFuIIx0YTEp/MYAmTViPH9rYQ\n7X7Jq4o/g1CX6qrmFNVkwMyUktKXCwLbr1YbPZ9SpOGNwLNfgHxatSowfEoUVkir6wggl4RiVtXF\nTJ9Upnt1V/zsDJx4XPVSAUVAm18N/ogiMX9E/Vye47PRU7QbdIj8hSyUl1GNlUDNMS4/PCuxbkaR\n7uoIzutkFjA1RpIFNjf7+dyjR+aJ2fq6oty8o5Uv/XTxIYkhr646skmV5dGFxNQgZy+9lqUCtZqY\n013t2RTh+HiaZNaiJaiaMMVCXhwp8U4OsHXmadJamLzw45VFjPwottDxOjl0p4DjD5Mv2nRljuCN\nxJSuZPKoshbsgmpHsONtyoWZOgavPAC2rZ6YwTY1SiMbV0SRT6gxpVJCfFB9XvdCIQH5kOojG+lW\nT9jHP3X+E9jfokgpXCo8DLUriwYBHddcOEVbj3jDWopZbNC09KIWipQyXPG9JqXUS6+VX+uCTAaG\n43x3YJzdnSFagibD8SzPn4nTEfbQ1xVZcDj6L7+uB/MCpZOpvCITTYCpC0xDr1lzshTKhBP2GWSL\nDtdvbWIkmZ+dGfR7b9/Fx2/dzU3T30ArJGnPnaRFJAmLDJq0ETh4sNkqRgkVJ9mVeppN2gyBLa9S\nF+6+QQVem3pUqjeXgDM/LlkrRdANdWz6lLJEvGEoJFWMBaEK7ApJMIKKcLxRRSyVbQ8WegJ3XqOa\nM1UXxe2/68LFcgtZO8sdrF6Pa9QTG7RAcDld7/1AEzAjpVxX4rbKgGxvLMRPTkwS9BbJWqq8fyEx\nW19XlKDXpJAtLnhNiQq8Fi1JzrJRodr6QdcgFvKQyBV59bYWTEPn//mFvXNvGOmHxBMgcmqglqOp\nBI1WcikiHeANERJ6qdVAULkx5ceEbsLMadUDReiKWMyA+qO2LdB19ZqZVIQR7lKK0h/fo4Rv/mbV\naEnaaoMKXQVYAy1KKDd9am6kRhmGT5GSv+n8VGxsx+Ip2nrEG9ZazGKDpqWXJBQhxJuAzwDXUwod\nCCGeB/5QSvlYg9dXF1QHZBO5ImGvTqoiplKpNC3rVXKWg0dX8vpqqTxQN4ukEpoAv6nhNXRSeZve\nsO+8WM/AcJzCN/6aK7MJJYTTvejSQpGaDoEmuPK2+eb05HEYK81qGz8IaKqNY8fVKiiKpkgi3KXI\nwJYqWJtLKrK45WNzm+CBX1OWST4BgZgioswUJM6qGEykG4r5+fVCZWHYYlmMC6Vo6yGDX4tS+g1Y\nIHhBQinN3vkWanzG7wNDQDfwXuAhIcQbpJTPNHyVl4hq8VrEZxLPFoksoDQt61Vs20HXVArXlso3\n1HVBsREsUgGPLrAcsPIWvc4pbso9w6Zzk6R9nTzh/UXadt7AfU+c5OPxl3A0D4ZTQNpFHAFaWXer\nmee7Gy3bIDkM556e8903lbrUT51QQjYrr9oTgOphYhUg2Apv/QvYc4c63rlHkVVln9lcAnJxiGyd\nqweqrhe6lCdwPeINGzRmsdawVIOljwGflVLeLaV8XEp5WEr5PSnlb6Gslo/Ve0FCiBYhxANCiLQQ\n4rQQ4pcu9Zq37umYnfvrSElnRHWg7wyr4Gb53K6OIH/24EGeOTnJk8cnCXt1nJJATQgImPqSGZqL\nRfmyAY+BISS9zinu0r9FlDSTegytkCD5+N/ytYcPqPnJmoYjDEQpnyRlqeEKKFKonuk7eUJZCGYA\nOvZAsH0uONt5jXKbpk/C6IAiB38rXPEmuPMrc2RSRtn/1z1Kbdtzk8r8bH3N/Pe1blftIu+4F970\nhxf/NC7fb+IYnHpCBY5P/VApdZd7jQ0Ws1hrWIpQXgf8j0XO/U8aM+Trv6PGc3QA/wn4OyHE1Zdy\nweph6L2xEH/wjt1sawvNBjrLitmpVAFHqpqdgq3cD0NTgVJbSowGEYoETA02NflAaNyqPUOcACkt\nhKHr5PQwCRmkffBRwj6D4fAeNFlECoEtTIqYKu5h+lWwtXrzjA+oDRiMKcvC9KmYxsThUuA1olwg\nXZ/rL7u3xOXVjaMX6sa2/U3qepWoxQKopTF15x6l1B1+AcYOq8CulPDil2oPqq6XBtbrHOJCPZOE\nEAkpZeQC55OV2aBLXowQQWAa2COlPFI6dj8wKKX8g8U+t2/fPvnss89e9H0HhuP82YMHmUoVyBQt\nUjmLkFfHdiBdsEq6D0FzwMvx8aRSw7J80dpSCJgCQ9cpWA7/lXsZppVWEuzShwjLNHHpJy5DfFT7\nGLe3T/GHiT/HsAsYdhYhJB5fEHpuBl8YXvsb8wN+Iy8pyyQ9oXQmuldZGNkpFaCNXTnnpsCcS1PM\nqGBmZYHchVSwtgWpIUhPKhXtLR9XFs5CKVuYU4sudf2HPqKsEm9EBZmtvIrh9N5cW/Wwi0uCEOI5\nKeW+pd5Xc5ZnEdR7T+0C7DKZlPAi8IbqNwoh7gLuAti6detF37AcM5lM5WkJmOQTFqm8RSZvIUrF\nffmiKsArWhKrFEMpq1frBUOAZUu2tPiYTOcZysboESPs1s6RlyZxAoRFhghZeu2T/OtQD2+O7qNP\nP4VwCjS3tOLZdPV8pWrlpnzoI3DssdL4C11lZ7JTymLxRpV7UglfBA4/otyZWjIjZSvih/9NWTiB\nVhUEPfKIOl9dB/Tje5T7VWvm5dwzKn1d7rRm+gA5F/h1sSawFKEEhRBnFjknUDU+9UQIiFcdizOX\n7JyFlPI+4D5QFsrF3vD+p05zYjxFPFtkOl3AciQeDfI2s3Spa5C3HMI+HV0IHCmXRSa7xWlu1Z6h\nW0wwKGMccPZzWPaoDI1QFccBj06u6DCdKSAQfFvu56+1e5EI8hh4KSCQnNR7+DnzeY4Vevl86g18\nuhM6OjoItrQtXmI/0q+CrvnEXLf3fFJlaGK71aZMjUJ0syoILBcMWrkLqzkXktb3vP78WqBn7psv\nXCu/nn5S1QEtdv1qVP+TX7YNSdculiKUle7GlgKqXawIkLyUiy7W4nFgOM6Pjk3S7DfojHg5OpaC\nklhNAJo211y6M+plU1MAXcswOJOr+d67xWnu0r9FnADDtBIVae7Sv8V99js5InvUWAwBWcsh6NXJ\nFmw0TWPYewVni21EyRAmSwI//bKXjNbCDvsEvyW+RhcTdMa20OIVKi6wWCZl4CEVHI1smsvKaIZq\noKTvVUK349+DiSPKgjADinyko9pDVmpJynGRhYrbTjwO26qMSV9EZZd6bjr/ePl6tWReNu9XLo8Q\ncy5PITk34dDFmsAFCUVK+YOVWkgJRwBDCLFTSnm0dGwv8MrFXvBCLR4P9I/iMzRGEnnscixJSLKl\nEhVTn2vhGM9aOE6aeLa4rAdjObiaIAgw+3qr9gwnZA+2o/ZI1GcQ9plMpgt0hL3MZAuckNsJOCmm\nHWUICqDHHmSTGOecaGNMtPLySJa9sSxNb/no0joOoc3Vypx8QonWyps50KLqcbJTKlW86Tq1uccH\nlPuyUOPopaT1oD4T7lqYODbvV9eDpZsY7f+wsrLSpYHvhhdarlDHl8JaktxvcFxqDKWukFKmhRD/\nBvxXIcSvoDrF/RyXkE060D+KbTscHE6oYKvPoDPs5UD/KAeH4xQsm7zl4DEEuiYolhRshiZmLRUB\nZAs2iariwlrQLSYYpnXesSQBNosJLFu5UwGPQUvQi6YJrt/axHiqQCJn8S25j981/oW9xDFlkTwm\nUVL0cwVxGWRz1If0ejmRSnP9hRSfC2kwMpOKOMpwbGjdoWp7yk/9YEwFZRdStv7k3vOFYp3XqDnH\n2Zn5BLH/rrlYSnXwFWpTi5b7wS6XGCotKc2Eo4/CS19VWamVnhh4GWBNEUoJvwH8b2AMmAR+XUp5\n0RbKK0Nxzk1l8ZoaIa9OvmhzZDRFpjhHEALJVNrCqqjaKzoSWVS9Thypfq4VlTGTLWIMUxY5Tdfs\n+TAZxvR2dEdgaJLWkIfmoIcd7Wqi4MBwgtFEFierArUI0ITyvXyiSE742Rz10dUUQErJRM57YcXn\nQu0MdRNCm+be44uoIkB/RZPrXAK6rlUakmosRFL5lJLwn35S/bx5/1zG5kLSepg7Vx64dalzfiqv\nW+7jUm694G9WTaA2QP+RtYY1RyhSyingjiXfWCMSOeW/lEd2+kydvOWQyFkIKRlP5SnaclYgVglL\ngpAXbkVQjeqYiSktXqUdAwfO0EmYDG1GlgORd3F9oInJdIF3XTu3sU+OpzDGD/Krzg+5Uf8pRQwG\n2MaM1oRlS96ovcjNoRFON6msTMFyiJl5iG5bfFEL1Y3c/DFlNZStifAmVUHcdqWKnSzVQ7WapCZP\nKAn/5v3zpfaVa6h3x/xaXJmyu3f6ybnWC1KqoHR5+qBLKHXDRROKEKIZuFNKeW8d11N3RP0G8VKP\nV6+hkbccpJTqeNbC0DQkDtnCfG1JWb9WdJTgrFbdSXXM5DSd4ECnmKIoTQZljG9ab+TQVIymbBrL\ngVMTKba2BknmLAqDL/Mh/d857hj4yNMi4vQywlnZzlFjJ4dkL2/Lv0TP1JMIK08ek5aObuj77fMX\ns9SGq7QaWq+AK981f7hXmUyq+8iW32MG1JM/MahUt2YAhn8GZ3+idC6+qBpZcSGdyKV0zK+FiMqW\nVHl0KsxND1ztWp4NiGURihBCB24DPgC8CzgKrGlCuaorSsDUGUnkSeSKRHwmva0BemMhDg7HcaTE\nZ+hkC848MhHMNZBWKlbV9X4pUlkoZnKGDorS4OPWr80e8+iqIdO13REOjygX7KquKFeLpxnKeXHs\nNGGRUf2LMGljBj+HiZttOJqXgu3g1zXag14C3gX+G2vZcNVWw0i/IosyJo7Bi19WgVArD4PPQ+p/\nQahNKWl1r4qz7P0lOP1j9XNmWkWZi1klcnvl31SryK5r5xPaYr1lobaNXisRlS0p3aNaXQqhWix0\n7XVreRqAmghFCHE9ikTuBPyAF/h5KeUKTpi+ONy6p4P7nsjQ1xWZbUIdzxa5dU8HAEdGkowk8ud9\nztAEhdLYDF0T+AyNQg3dkwZljKhIz1oooGImgzI2730FGxwp6W4J0NsWJuo32dURxEoPcdZqZp8Y\nYooIrTIBAnTUTNIrOUFwx5tpqVa1Vm+kgYfUhh59WW0cX0R1a6t1WNfEMXj675UF4o2qZkvJYZVO\nFqgm1lZeFRb+6G9UhicxpGIzuqEIJTWirIL8zHxCg4V7y255jSKVWjZ6rdXDZXfvmS+otLa/RQ01\n0z3uWIwGYKlq498D/jOwE/gO8DvAg8Bx4OmGr64OKNfxVOpQ3rd/82zfk2/+bBCvruHRBcm8IgwJ\nFJy5nrAFS5K3aut3csDZz136twCVzQmTIUqGrzpvnPc+XahWjk8emyTk0ZnKFMlbDr8mYoRJ0sYM\nHoroOPjIYwuTJAG2mvmFVa3VG+nkj1QfWClVfY9tQXZaZW0WQuUTPzWmPlvMAkLpPbJTqm2kMNQ1\nhJhTq06dgKvuUCM+jVKMophVLSILKTj3LMychdbdc0HXhXrLjh+qfaMvp3q4PNyr0gX0d26I/iNr\nDUtZKJ9FZVo+APxreVi6EA2qkGsQqptQVx7f0R5i2JtlNJHHdiS5onPerJzlBGUPyx7us985Txn7\nVeeNHJY957037DMYjecYlaq7pu04HDBfzd36l2mSKTXbGAsTmxwe/IEwnljz0mKwkX6YPKwauXgC\nJZn9BHgikK8WIpdQ+cSfOKzcFwCnoERwmqGsFVlUepbKfxzdo4ikuReSo+p9VkHdVxqgldyNwWeU\n+ralZ+5e5d6yg8+qfi35pArsLoWLGcS1AfuPrDXUopT9AKofyueEEF8BvsQGEj1fvSnK5ubAbK+U\no6MJHukfnes+fxHXPCx7OGyfTyCVCHkNpjN5NbIUid/UEULQX9zMqNlCrwiyiQmK6KQI4REOO4oD\nsPNjMFbKoi+2kQYeUgFSmVb6Ek0HO6d6k9h5FWStjmdMn1KWRDCmLJRgmyIOKcpttZRVIuXca1mt\nuvVGtYbWnWo9mSlFPKDu59hKdiyB+GnY9vrzrYtiVgnVdrxFqXMf+FWIbj0/9lLGBu14tt6xlFL2\n+8D3hRC/Cfw8ilx+B/Un9qtCiHullJMNX2UDoWIsJ4Fy17YcosQkugZWtblSJ6Ry1mzTSFMTSAmG\npmFZFppVYNxoJidNwuSIGDZeXwjDNFXPEjMwX+ux6x1qY/3kXrWxhl+aC8QWksrtKOZUbKN568Lx\njPAmFVDNxpVrJB1llRgo4nBUDAfNVHGSSrXqG0ptcZ75goqxWJUxKQk4qoObdJSbVG1dDL+ovu/a\nq6qhxw6qe1XHXhYiFZdA1hRqncuTBe4H7hdCbEYRy/uBP6T+BYIrinKM5f6nTvPdgUmGZnIYQlCU\ncnbIVyNQeWnbkWy1TvI2fkq3psRwnUwyKmLYZhNNsSAeYamYxInHYefbVVFdLqEI5sUvl2p1SiQS\nP6MCsLm4aqQkpbIwdP38uTcwF8/whpW7k08qYgjGlAvjOKUVa8o6CXfBpr3np6JD7aoXC6hgsF1S\nFktbVTibfkVC1daFXVC1PqF2Va9j+Ep9cJOr3/vVxbKwbB2KlPIc8FfAXwkhXrPU+9c6Bobj3PPd\no/zo+CS6JvDogqIusKVsSM9YQ1P7u/LaO8VpPiSUGG6IVnyaxdWcxWsmyZitxJMpgkFRGm/RMj9V\nem5Cfb/purljbX0w+JzamPEzyurw+GHrzQunZivjGaF2ZUmM9MPoQbCyzE5L0XRlqRQyqt/KQiIy\nO696lujeklVT+ri01DlfKaVeaV08/ilFhKCqnos55QJ5AiX3K+bqRdYJlsry3A68S0r5qwuc+3vg\nm41a2EpgYDjOZw8c5oWzM5i6QBeCbNFGIjFK7VnLM41BDSBaTm/7Sgn+kIzxbbl/tsK4EpViuFZm\n6DESSC1A1JrAFDaTeptK0w49B51VHG7nzw/0eILKysgllMsjbbCquvdXBnLL8YzUmLJQ0iWSSpxj\nVuInpLoOmhKy/eO7VIwj0j0X54hugbFDyuXxN5WyPenS5w01fB1rrutbGWUXKDMJmVKXOU0DM6TS\nye1Xza94drFmsVQLyI8CX1zk3BdpQE/ZlcSB/lGm0gUEKoVrGhqmrqnWrEJAKf5oaOA3BQGvPhu8\nXQplCX6LnmG0om3BLk6f995uMUGSAK3McIN2FMPJE/d2kzFbyGkBpC+qVK0LtVnUvco9qMSJHyiX\nJZ9QG9kTUinf448rkVllP9XKfq1nf6KsGbsIiWG1sUG5K5QCsbIISNU2cvqUEsJNHFOEUG4xmU+A\nGVT3pCSAM/2K/MpamGqYAVVY6BQBR5FPoFndd3zA7f26TrCUy3OVlPKHi5x7ErikXq+rjVeG4gxO\nZ8kVbQqWjalr5IoOUoLH0PB7NPJFm6BHDfHShODKjhA/PjG1ZPbnVu0ZEgRIOAEMXZCwg2jAf9Ce\nYaAqAzQkY/SKYW7QjhAmi3Akekojr4cY9Paydfs1qkCvLD6DuQxPMAaIuZqcqZMly0JTwVNEKUgq\noJiCw99SEvvKIOeNd8MjH1cuSrC5NOgrrtLBdq7U/LriNxa6SgV7Aso9SQ0ronjmPiWCMwPKMhG6\nMut0b2msaRiSgzBcQcqVgrpQu7pubgZMr7Ky/BHwlly8yhIAtwXBmsRShOIXQoSllAs1OAqhVLPr\nEgPDcc5Nqzk8flMjXbDJFS10DXRNw3Ikm5q8zGSUq3BFLMhUOs+h0VRNqeSyBF+g3Kaw1yBtBdnC\nJLqtQpzl6xxmM+/RfkSEDIawEEKgYzFjC3ZxhkiqZIEsmCr9ROkXekhld0ZfVj87VmmIF6WnPnOV\nttXits49Kqjbc5NKFR96WMVcnFLco/o3lrYiG6GDk1euUnZaEVLPTdC0VZGC55AKzFZWMGfj87Uw\nlYI6X1QRlL9ZZYN6b57rbXuxBYT1hNtXZUksRSgvoNLF/98C594L/KzuK1ohHOgfpTPiZXgmS6Zo\nz7Z2lDb4PTrhkntj6oIzU8qKSectirazpD5FMCfBz+qhksWj06xn8YZ6uSXcRjpvcWQshXQcrnYG\ned7ZwZv1F9AF2BhYhoegkORsiOQqNuBiqdKJY3Do39Vm0wxwciUyKFUmSQeMgLJoFqqyrVSe+iKK\nTNITKk3sFKpuJlUqWtPUeStXqutpVp8px2HSY+BrVi5ZucuadBRxlFEpqIvtLjXQ9pTS1zNzoz9W\ne+rfpVRFX0ZYKobyV8DfCiE+KoToEUJ4Sq8fBf4W+GTjl9gYHByOMxzP0RHx0VSKiwgUmdx2TSeb\nmgPkLQdd0+hpCeAzdZI5G48uWEoorAl4jP00iwxBJ10ql5/Bbyc52vpGPn7rbv7i567mxm0tSARb\ntElGjS7iook8XnQkppUlWCxJ5b3nq3znYaQfnvisYjl/ac6w0EpKVUt9gYp7TB6DY99VVkjlCIrK\nuTWtOxWhiAuMrnYKyiUxfYCmiCLQpoKoxZzKRqGp6+ST6r7lTE2ooqNbdMvcDKGyapaSiK486sLO\nX7i37Uqg0pISWol4mxaOB13GuCChSCm/DXwYJWY7CWSBE8BvA78ipfxOw1fYIMSzyrVoDnrY3hZi\n75YmemKKODyGTrPf4PRUhuNjKWRpOFgs5MFr6izBJwC8YisJflwG2KxPIb1NvND9y5wyVB1OX1eU\ne9+/j6/+6mtp37KTTk+BnBbAxFH/KwIcoeMR9vyWigth4CG1cX0R1XhalKyScjC1DN1QmpLMpHp/\neVh42ZTPJZTLlB5T7kZruQBRV0FWM6iCvABoynJo64OmzWqkaWZ8rueIXVBNr6VUja+jmyGyWblf\niaE5MqsewKV7ILYT3vP3c8PBKkmnjJWuFF5o+Lvb/uA8LKlDkVJ+DfiaEGI30ApMSikPN3xlDUbE\nZ5DIFOf1SfHqOjvbvRQsm4GRFG0hD15DI1eUHB5JEfUbnJrKzg7oWwxljckhp4dj9BIUOj3hIN6i\nzlV+87yh7Nz2QWYe+xsKJ3PojoXmgCMECb2ZqL+GMFX8rGrnWMyrlLHhAcerUsWGrxTrKM6pX6Wt\nXBBfk1K3lmfvdO6paNv4YfX1D7ep60Ip5eUBPap46tX/1/xgcXpCWSbFkru1+dVKDxPPqMyRL6L0\nMrpnzl2pRUJ/MXU79YY7yrQmLKVDCQB/AuwBngc+JaU8v9Z/HeLqTaU+Kck8E8kcOctBQxAJKPfn\nzVe2z0sRn5pI8f3D47SHPZiaYDxVHVdQ0IUqntSFEsYZusByJF5DI5Wz5g1ln0XnHpr2vhvr9Hcp\naia2lAih0WzkMbv2l2IhVagMEE6fAj0w9xR3HDD8yoLwRhXJTB5Vm1o3lGtRSKnhX2MHlXy+a++c\nKQ/q2m/6Q9j2RtUR386peIlRCtYGYnOp3DIpPPJxRSrBmLpeqF3FV9p2w7Zb5tYunflP9qUk9Guh\nbmctkNo6wFIWyueB/cAjqOBsK3B3oxe1Eij3SekMe0lkingMHSklXREvPzo2yY1XtMwjlOF4Dk0I\nAl6TlpCPjojFqYkU6VJjJp25zI0uSpkiy8GR4NUEecshVOrH0t20gNUxPoAR26H+Q8rDrIo5lQLu\neNv891YHCK1Sv9SWHWBnSvoRqaYIZibUdbwlq6Jpq6rZSU+qwKeVVa0fczNqBEaofb4p/4aPKVKY\nOa3IwsorCX+4Qk9SJn8SfpwAABPwSURBVIR3fHb+JMCJY6pQ0MopSX1sd+39Tqqx2nU7a4HU1gGW\nIpR3ANdLKYeFEPcAT7BBCKVcw/NnDx6k6DjEQl52tAeJhXycmswwMJykIzK38SfTBTY1+ciXqgUD\nHp2A18BBtZH06Bo5yyZXtLFsqdqQ6AJDU0V/+aJDT0uAeLbI+/ZvPn9B8bOqa/xgaaSq4QWkSsdW\ni7qqu5WVVaTJIWjepup1EkNq0/ubVSGhZih3JRtX5xxLuTDeaKm9wTQMPQ+7bp2/4Tv3wN47lcYk\nn1Sk1HkdRLsX7iBf3nTDLynZf9d1SnuSjcOZn6hYi6avzyf7apPaOsCSkwOllMMAUsqzQogl0g3r\nC31dUba2BLiiLcCJ8QzPn5kh4jPZFPXyyrCaJlju8mbqGle0hQj5DI6Np0nlLAqWJGDqNAc9TKWL\neDSBrWtYOPhNja4mP4mcRWvQQ2fUx7a20OyQsfNQ8tGnWvaSHhxAy48hTC/eLTfSuljj5Uq0bFOE\ncUepI2elS1Qei5EcURqQmdPgDanh6KCICE1ZKuVUbXnDj/SrZtYd16AILqEaKU2fKFkii3SQP/VD\n5WJZGZU1ykwoCyc5pCyZemxMVxey5rAUoRhCiDcxlyqo/hkp5fcatbiVgFcXPHViClMXJHMWI/Ec\nUsKebtWWsdzl7e43X8F3B8bxGDrbYwEGhpNIJHlL4tE1uqI+pjMFik4RXdMJ+kxu3dOlCEScnfvD\nP7QFxAJ/+H23M/PY3/DyBEj/tYQDWfRCnO9k38Ltw/H5JKR75+YU+yLKldA9892ICz1N/+71Kh7i\nKVtgm1RRXrkGp9KUr7SGckl1v8yk+gsId57fQR7OD9BOHlPtHXtuUnVA9SITVxey5rAUoYyhZuSU\nMVn1swSq+hGuL6gWjw5TaQuPLjB1Qa7ocG4qyx/fNt+a2N4W4v6nTvPj41M0B0xu3N7CUyemOTud\nZUuzn1jIS9hnsqs9xLa2EB95267a//A79/Bg4D+yxftd2pwxkp4ujnb8HBm9d15W6Lw5xcWssgZa\nts+pZpdC9VhPzVCy+N3vPH8GT6U15IsogpDWXP+F6g7yZQIKxtR7y/GgicPnk96l4GK75btoKJZq\nsNS7QutYcZTnHT95fJJssagGegFeXaMt5CVvyfkbGeUitYV98zJAQgh+fHyKoZkcOztC9LQE0HVt\ntgl2zX/4I/10DH2HGBMkfV0cbXkTk6GdhKWcnxVaaE6xN6I2fa0baTljPaNbVHA1Nawk9rkZ5coY\nPkUYVm5+B/kyAZVVr6CIJD1R36xIrU2qXawo1tygr5VA5bzjjrCX6XQBU4euqI+AxyBXtAl6tfPT\nu8DgTJau6FzF786OCNGAyYtnE2xtDc4bxg7U9odfsmK2OFP4c6N0pfq5YuJxfrL5wzwffsNcVmik\nXylckUq+Xs6aSEe5ErViOWM92/rgxa+oiuVArHSvEUUSoKydysbSAw8pK6ysei3L8IOx+rojtepC\n3DjLiuKyJJQD/aNE/SZRv8mO9hDHx9PYjmQylUcLqxRvz//f3rkHx3VXd/xzvJK1siVLseU3sUxC\nbJIYEoLdpk3wmDrBAcYhNJ0WkhZoh8kfTKZDW5rM0Jo6lCHttNN2eHXGZCiQlpIHSYgLKe10YqhT\nHjEw9cT1Y/JyHMtyJCdWJFuybPn0j3Nv9mq1b997d1d7PjN3rL0P7dmf9fvu+Z3f+Z3fwnkFp3dX\n9nYyMn522pRyR1sbN16x1IY4+VTyhx9seXGJvsTgORjL9NCpp9lw5D4OLFvKTVtuyA2dMnOD7NNX\n4cBOm6XJ9piXUAn5HaxQoaQoQ/tNNEYHzO65XXBRvyXKtWfNa1n+9ulxlzBfY35fTmwqFZNKBaCS\nvBCPs6ROubU8s5KjJ8fpzpqWLu7O8iurLyIzRxidsKzZNUu6pg9bIty0bikj42cZGT/LedU3fi50\nLzAztTxaiyRk5AiMDZDNzmfxwh4ybRnG6KSdKT62aK95O+HQaflVNlMzdtyEZXLM4imvH52+NqcQ\nYQcbPzm9g5V6buSIzSCtfpcJi06ZkGQ6bOYnu2B6pw+njjt7zaZwPU6lYlLKvsFnrITBY5+w9ljz\n3tLvk8T6m6gNT95bvs1bjJb0UPK9jDXLFtCemcPg6BlWBZ5Jsendcvv8zKCShKiei63i/LxFzBdh\n/tw2i0+0rYCpIbsnHDrJHAugTp62XBI9b509ms5ejELxnNMnLMP1otWFPYKohxXdXqOzp3g8qNZ8\njVLxJpjpbRx6orRYxR1ncY+nLC0pKPmV7kcnzpHJzOGem68oLgwRiu3zU5RyHezyrbnFedkFNnNy\nbsJmbsKhUbRjn58yr+HcGfMWwjhKuY6S38HGXoHj+8zr6L+ucAcJhxanhmH4kNW1nZOx+yHeQGgp\nAahlVifu9Tc+s1SWlhzyhF5GT2c7x0Ym6Ols546Nb65OJOJk2TrYeJfldpw+YbMuS66wjhsOjaJD\np45u6xhTZywwC5V1lPxVu8MHzeOZ32eC8dKPLfv1/g/Czj/KufNnx63kwZmx4P7FtjZo7JV4F8iV\nWlVcy2rfSoab1eArjsvSkh4K1OBlJM26W6xubLGAZHTo1NlrJRr7ghqu+dmtxcgPZJ4aNtGa12f1\nXMdfs6HTuUnLUxk6YOJ1ati2P50cD7JqFTJZy5DtuwxW/Vrh8ozVzrCUCrSGs0fVeBtxr7/xFcdl\nES23Fr8JWL9+ve7Zs6feZqRLrdOh+auUu5ZbjsmJ53NbJWbabD+f0QHbAGxq0oRFBE69arVpsz2A\nwLs+ZbGMcEFgKAJr3lv8/ND+4nYX+1zR+EX09xWLXyQxXVytDbMIEfm5qq4ve58LSguR38kWX26d\nfuCXNg0tGYun9Ky00gfH90Hvxblar+3ZXKp9/3W5b+r8b+3xk1aoaenbpp8fftY2Ye+/vrYOWalI\nJNnxWzSvpVJBadkhT8tRaIbi0BPmMYwcsWFNpg26Vlr9lLMTVtU+0zE961V1eiLbT75SOJA6eiwX\nuA0ZG7As21qDmpXOHiUZPPUVxyVpyaBsS1IsJ2Nov63+fdN628JidBCO77ctTucvzSWnrQy+nCZe\ns1yY8Nu+WCC1e/nM86dOwLxF088lEdT04GndcA8lZcI1RGEOS9FyBnFTakp22Tq46jb4r3tg8qwV\nUJrXB91L7XwY87jsxpkufrFA6oY7zAOKnp/TNtOGJIKaHjytGy4oKRJdQ7S8J8vI+Fl2/OiFdKas\ny3Wyof1w2XtmxkKG9s9cgRxl2TobNj29w4Y53ctNTArNWm28y0Qm3JQsqTKKXq6xbrigpEh0DRHw\nxr/5q5oToVwnqzWrNFqAqf86+72HnjAxKRRvyBeZJMooernGuuGCkiL5K5WBwkWrk6BcJ6t1mFBt\nADStoKYHT+tCwwiKiOwCrgWCXak4qqpr62dR/BRaqVy0aHUSlOpktQ4TvC6JE6HRZnnuVNWu4JhV\nYgI1rFROk1pXCYflKA98z7Jr407Hd5qKhvFQWoGqVyqnTbXDhMFnTHzOvB6sgA7LUV4K12xLxsYW\nTSxrFhomUzYY8lyJJYAfBP5MVXeVuP8O4A6AVatWvfPw4cMpWOlM48l7Le4yNZkrR5mZa3kqW/8+\n/vdr4dT3etOMmbJ3A/8HTAIfAnaKyNWq+lyhm1V1B7ADLPU+NSudHNEaLV1L7Fy15SirwcsHNDyp\nxFBEZJeIaJFjN4Cq/lRVR1X1jKp+A3gKeF8a9jk1kvYm5sf2wvG90+M1HgBuKFLxUFR1Uy2PEdn/\nx2lA0kwgG3zGdiJUctt5vPwz2yWx7y21/06Px8RKQ8zyiEiviGwRkayItInI7cBG4Af1ts0pwYXU\nj62W/TttdTRqleraOkxchvbXVjCplvq6TlkaJYbSDnwOeCswBRwAblHVg3W1yilPWglkYbHsju5c\nADi7wEQs7vq17qXUTEMIiqoOARvqbYfTwISZvF1LcgHg/MzeavCEvERoCEFxnBkUKwYF8cRrfEVy\nIjREDMVxplEovhEWg4orXhN3AWsHcA/FaUSKxTfKlVKoBl+RnAguKE7jkVZ8w1ckx44PeZzGI+2E\nOSc2XFCcxsPjG02LC4rTeKSZMOfEisdQnHSoNs3d4xtNiXsoTvJ4mnvL4ILiJE+xPYH276y3ZU7M\nuKA4yeMbb7UMLihO8vg0cMvgguIkj08DtwwuKE7y+DRwy+DTxk46+DRwS+CC4jhJ0mJlJn3I4zhJ\n0YL5N+6hOE5SNGKZyYQ9JvdQHCcpGi3/JgWPyQXFcZKi0fJvUshYdkFxnKRotPybFDwmFxTHSYpG\ny79JwWPyoKzjJEkj5d+ksNOjeyiO0yqk4DG5h+I4rUTCHpN7KI7jxIYLiuM4seGC4jhObLigOI4T\nGy4ojuPEhguK4zix4YLiOE5suKA4jhMboqr1tuGCEZEh4HCNj/cBwzGakzTNZi+4zWmQtL39qrq4\n3E2zQlAuBBHZo6rr621HpTSbveA2p0Gj2OtDHsdxYsMFxXGc2HBBgR31NqBKms1ecJvToCHsbfkY\niuM48eEeiuM4seGC4jhObLigACKyS0QmRGQsOA7W26Z8RGShiDwqIqdE5LCI3FZvm8rRJO16p4js\nEZEzIvL1vGubReSAiJwWkSdFpL9OZkZtKmiviKwWEY209ZiIbEvbPheUHHeqaldwrK23MQX4MjAJ\nLAVuB/5RRK6sr0kV0ejtOgB8Dvha9KSI9AGPANuAhcAe4IHUrZtJQXsj9Eba+y9TtAtwQWkKRGQ+\ncCuwTVXHVHU38DgQX3XhFkVVH1HVx4ATeZd+E9inqg+p6gSwHbhKRN6ato1RStjbELig5LhXRIZF\n5CkR2VRvY/JYA0yp6qHIuf8FmsFDaeR2LcWVWBsDoKqngOdo/DY/LCIvi8g/BV5WqrigGHcDlwAr\nsfn8nSJyaX1NmkYXMJJ3bgToroMt1dDo7VqKZmvzYWAD0A+8E7PzX9I2YtYLShAY1CLHbgBV/amq\njqrqGVX9BvAU8L76Wj6NMSBvyzcWAKN1sKVimqBdS9FUbR4Mhfeo6jlVPQ7cCbxHRPI/Q6LMekFR\n1U2qKkWO64s9BkiadpbhENAmIpdFzl0F7KuTPbXSaO1ain1YGwNvxLEupXnaPMxYTbW9Z72glENE\nekVki4hkRaRNRG4HNgI/qLdtIcH4/RHgsyIyX0SuAz4A3F9fy4rTDO0KENiWBTJAJrQXeBRYJyK3\nBtc/A+xV1QONaK+I/KqIrBWROSKyCPgCsEtV84dtyaKqLX0Ai4GnMVf2JPAT4MZ621XAzoXAY8Ap\n4CXgtnrbNEvadTv2bR49tgfXbgAOAOPALmB1o9oLfBh4Ifj7OAZ8E1iWtn2+lsdxnNho+SGP4zjx\n4YLiOE5suKA4jhMbLiiO48SGC4rjOLHhguI4Tmy4oDgXRFB345J62xEiIp8WkfvqbUer4oLSRIjI\niyIyHnTi48GK0q7I9S0i8iMRGRWRIRH5oYjcnPc7NgXrmO4q816bROR8pFjPyyLyoIhsiN6nVnfj\n+Xg/ae2o6udV9eNx/14RWS4ij4vIQNB+q+N+j9mAC0rzsVVVu4BrsNWlfw4gIr8FPIRlSL4JK8T0\nGWBr3vMfBV4N/i3HQPBe3cC1WNbof4vI5hg+R7NxHvh3rC6NU4x6pxL7UfkBvAjcEHn9N8C/YQvA\nXgL+tMzz87BU+A9h1d/Wl7h3E/BygfNfAvZEXivwluDnrwNfAZ7AVus+BSwD/gF4DROkd0SeXQF8\nBxjC0sb/MHJtO/AgJpCj2KK89ZHrdwNHg2sHgc2R5/45ct/NwbMnsfT5y/Pa81PAXqw0wQNAtkwb\ntgWfeXW9/x4a8XAPpUkRkYuxUgC/BNYCFwMPl3nsVqyjP4Qt0vtIDW/9CHBNsPq2EL+NeU19wBng\nx8AvgtcPA38X2D8H2IkVMVoJbAY+KSJbIr/rZuDbQC9Woe5LwbNrseX5G1S1G9iCicM0RGQN8K/A\nJ7G1Rd/HarLMzbP3JuDNwNuBj1XaEM5MXFCaj8dE5CSwG/gh8HlgUXDtWJlnPwo8oKpTwLeAD4tI\ne5XvP4B5RL1Frj+qqj9XK5v4KDChqt8M3vMB4B3BfRuAxar6WVWdVIvDfBXznkJ2q+r3g2fvJ1dO\nYAroAK4QkXZVfVFVnytgy+8A31PV/1TVs8DfAp3Ar0fu+YKqDqjqq5jAXV1VazjTcEFpPm5R1V5V\n7VfVT6jqOLn6osuLPRR4NO8mV8Xru0AWeH+V778Sc/lPFrl+PPLzeIHXYRC5H1ghIifDA/g0FvsJ\nGYz8fBrIikibqj6LeR3bgVdE5NsisqKALSuAw+ELVT0PHAk+Q7H36MKpGReU2cFBrKOUChj+Hvb/\nvVNEBoHnMUGpdtjzQeAXajVaLoQjwAuBOIZHt6pWVNFNVb+lViCrHxO4vy5w20BwHQAREWxoePQC\nbXeK4IIyC1CLFv4xsE1Efl9EFgSFdq4XkXDP248A92AufXjcCrw/KMhTFDFWishfAB/HPIkL5WfA\n6yJyt4h0ikhGRNblT0sXsWetiPyGiHQAE5jnM1Xg1gexz7c5GNr9CRbX+Z9aDA4KG3UELzuC104E\nF5RZgqo+jMUM/gD7Zj6O7d/yXRG5FlgNfFlVByPH48CzWHGeQqwQkTEskPs08DZgk6r+Rwz2TmFT\n2ldjMzzDwH1ATwWPdwB/FTwzCCyhgMip6kHgd4EvBvduxabdJ2s0exxrC8gVXnIieIElx3Fiwz0U\nx3FiwwXFcZzYcEFxHCc2XFAcx4kNFxTHcWLDBcVxnNhwQXEcJzZcUBzHiQ0XFMdxYuP/AWRax33J\n281qAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11add8908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# do and plot PCA\n",
    "# http://scikit-learn.org/stable/auto_examples/decomposition/plot_pca_vs_lda.html\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "pca = PCA(n_components=10).fit(X_scaled)\n",
    "Xp = pca.transform(X_scaled)\n",
    "\n",
    "plt.figure(figsize=(4,4))\n",
    "for i, label in enumerate(labels):\n",
    "    plt.scatter(Xp[Y==i, 0], Xp[Y == i, 1], alpha=.5,\n",
    "                label=label)\n",
    "\n",
    "plt.xlabel('PCA Dimension 1')\n",
    "plt.ylabel('PCA Dimension 2')\n",
    "plt.rcParams.update({'font.size': 12})\n",
    "plt.tight_layout()\n",
    "plt.legend()\n",
    "plt.savefig(\"figs/pca_1.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(1, 0.44),\n",
       " (2, 0.63),\n",
       " (3, 0.73),\n",
       " (4, 0.79),\n",
       " (5, 0.85),\n",
       " (6, 0.89),\n",
       " (7, 0.91),\n",
       " (8, 0.93),\n",
       " (9, 0.94),\n",
       " (10, 0.95)]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARQAAAEUCAYAAADqcMl5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGksoxpiksYRijEkaSyjGmKT5f2UFSXlXCdAoAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b31ba58>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(4,4))\n",
    "plt.plot(np.arange(1,11), pca.explained_variance_ratio_)\n",
    "plt.plot(np.arange(1,11), pca.explained_variance_ratio_, '.')\n",
    "\n",
    "pca.explained_variance_[0:4]\n",
    "plt.xlabel(r'PCA Components')\n",
    "plt.ylabel('Explained Variance')\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"figs/pca_2.pdf\")\n",
    "\n",
    "# cumulatively explained variance\n",
    "list(zip(np.arange(1,10+1),np.cumsum(pca.explained_variance_ratio_)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6 components explain 89 percent of the variance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAD/CAYAAAD8MdEiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nw4GAYRhmyPk/RoxjhWzfjxkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x118f5ee48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# lookat individual feature weights of first PCA component\n",
    "max_vec = pca.components_[0,:] \n",
    "plt.plot(max_vec)\n",
    "max_props = np.argsort(max_vec)[::-1][0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('concavity_mean', 7),\n",
       " ('compactness_mean', 6),\n",
       " ('concavity_worst', 27),\n",
       " ('smoothness_mean', 5),\n",
       " ('texture_worst', 22),\n",
       " ('compactness_worst', 26),\n",
       " ('fractal_dimension_se', 20),\n",
       " ('texture_mean', 2),\n",
       " ('perimeter_worst', 23),\n",
       " ('perimeter_mean', 3)]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# list components with maximal contribution to first component\n",
    "[(list(df)[prop],prop) for prop in max_props]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10, 30)\n",
      "[[ 0.26  0.26]\n",
      " [-0.03  0.06]]\n"
     ]
    }
   ],
   "source": [
    "# get component vector for plot\n",
    "V = pca.components_[0:2, [max_props[0],max_props[1]]]\n",
    "print(pca.components_.shape)\n",
    "print(V)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.00"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check orthogonality of components (over all 30 dimensions)\n",
    "np.dot(pca.components_[1,:], pca.components_[0,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARQAAAEUCAYAAADqcMl5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n18wEPFcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b53d9e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show direction of pca projection\n",
    "\n",
    "plt.figure(figsize=(4,4))\n",
    "for i, label in enumerate(labels):\n",
    "    plt.scatter(X_scaled[Y==i, max_props[0]], X_scaled[Y == i, max_props[1]], alpha=.5,\n",
    "                label=label)\n",
    "\n",
    "# scale arrows for visibility\n",
    "s = 3\n",
    "plt.arrow(0, 0, s*V[0, 0], s*V[0, 1] , fc='k', ec='k', width = .02, alpha=0.6)\n",
    "plt.arrow(0, 0, s*V[1, 0], s*V[1, 1] , fc='k', ec='k', width = .02, alpha=0.6)\n",
    "plt.xlabel('Original Dimension {}'.format(max_props[0]))\n",
    "plt.ylabel('Original Dimension {}'.format(max_props[1]))\n",
    "plt.legend(loc='best')\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"figs/pca_3.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Nearest Neighbor Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import datasets, neighbors, svm, preprocessing\n",
    "from sklearn.model_selection import cross_validate, KFold\n",
    "from sklearn.pipeline import make_pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 10 fold cross validation\n",
    "# see http://scikit-learn.org/stable/modules/cross_validation.html\n",
    "# and http://scikit-learn.org/stable/modules/model_evaluation.html\n",
    "score_list = ['accuracy', 'precision', 'recall']\n",
    "\n",
    "def cross_validate_with_statistics(classifier, X, Y):\n",
    "\n",
    "    scores = cross_validate(classifier, X, Y, \n",
    "                            cv=KFold(10, shuffle=True, random_state=42), \n",
    "                            scoring=score_list, return_train_score=True,\n",
    "                            )\n",
    "    \n",
    "    # get mean and std of \n",
    "    means = {s:np.mean(scores[s]) for s in scores}\n",
    "    stds = {s:np.std(scores[s]) for s in scores}\n",
    "\n",
    "    return means, stds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'fit_time': 0.00,\n",
       "  'score_time': 0.00,\n",
       "  'test_accuracy': 0.94,\n",
       "  'test_precision': 0.94,\n",
       "  'test_recall': 0.94,\n",
       "  'train_accuracy': 1.00,\n",
       "  'train_precision': 1.00,\n",
       "  'train_recall': 1.00},\n",
       " {'fit_time': 0.00,\n",
       "  'score_time': 0.00,\n",
       "  'test_accuracy': 0.03,\n",
       "  'test_precision': 0.05,\n",
       "  'test_recall': 0.06,\n",
       "  'train_accuracy': 0.00,\n",
       "  'train_precision': 0.00,\n",
       "  'train_recall': 0.00})"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def nearest_neighbor(X, Y):\n",
    "    # instantiate nearest neighbor classifier\n",
    "    nn = neighbors.KNeighborsClassifier(n_neighbors=1, algorithm='brute')\n",
    "    # standardise data to 0 mean and 1 variance\n",
    "    # note that this shows correct behaviour with test data\n",
    "    # i.e. test data is standardised w/ mean and variance of training data\n",
    "    nn = make_pipeline(preprocessing.StandardScaler(), nn)\n",
    "    nn_means, nn_stds = cross_validate_with_statistics(nn, X, Y)\n",
    "\n",
    "    return nn_means, nn_stds\n",
    "\n",
    "nearest_neighbor(X_balanced, Y_balanced)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "As always with NN, training error is 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# now do k nearest neighbour\n",
    "def k_nearest_neighbor(X, Y, ks):\n",
    "    \n",
    "    # for each k, save [accuracy, precision, and recall] as mean and std\n",
    "    scores_knn = np.zeros((len(ks), 3, 2))\n",
    "    for i, k in enumerate(ks):\n",
    "        # instantiate nearest neighbor classifier\n",
    "        knn = neighbors.KNeighborsClassifier(n_neighbors=k, algorithm='brute',\n",
    "                                            metric='euclidean',\n",
    "                                            weights='uniform')\n",
    "        # standardise data to 0 mean and 1 variance\n",
    "        knn = make_pipeline(preprocessing.StandardScaler(), knn)\n",
    "        means_knn, stds_knn = cross_validate_with_statistics(knn, X, Y)\n",
    "    \n",
    "        scores_knn[i, 0, 0] = means_knn['test_accuracy']\n",
    "        scores_knn[i, 1, 0] = means_knn['test_precision']\n",
    "        scores_knn[i, 2, 0] = means_knn['test_recall'] \n",
    "\n",
    "        scores_knn[i, 0, 1] = stds_knn['test_accuracy']\n",
    "        scores_knn[i, 1, 1] = stds_knn['test_precision']\n",
    "        scores_knn[i, 2, 1] = stds_knn['test_recall']\n",
    "    \n",
    "    return scores_knn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create plots of k-NN for report\n",
    "def knn_plot(knn_res, ks, s):\n",
    "    \n",
    "    from matplotlib.ticker import FormatStrFormatter\n",
    "    \n",
    "    # precision recall plot\n",
    "    # recall\n",
    "    xpl = knn_res[:,2,0]\n",
    "    # precision\n",
    "    ypl = knn_res[:,1,0]\n",
    "\n",
    "    xerr = knn_res[:,2,1]\n",
    "    yerr = knn_res[:,1,1]\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.yaxis.set_major_formatter(FormatStrFormatter('%.2f'))\n",
    "\n",
    "    plt.errorbar(xpl, ypl, yerr, xerr, fmt='none', alpha=0.3, elinewidth=0.5)\n",
    "\n",
    "    plt.scatter(xpl, ypl, c=ks, label='k-NN')\n",
    "    plt.colorbar(format=FormatStrFormatter('%i'), label='$k$')\n",
    "\n",
    "    plt.plot([0,0.9999,1],[1,1,0], label='ideal', c='gray', alpha=0.8, linewidth=0.5)\n",
    "\n",
    "    plt.annotate(r\"$k={}$\".format(ks[0]), (xpl[0]-0.02, ypl[0]-0.009))\n",
    "    plt.annotate(r\"$k={}$\".format(ks[-1]), (xpl[-1]-0.02, ypl[-1]-0.009))\n",
    "\n",
    "    plt.ylabel('Precision')\n",
    "    plt.xlabel('Recall')\n",
    "    plt.ylim(0.9, 1.005)\n",
    "    plt.xlim(0.4, 1.01)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig('figs/'+s+'.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "k, accuracy, precision, recall\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(1, array([ 0.95,  0.93,  0.91])),\n",
       " (2, array([ 0.95,  0.99,  0.88])),\n",
       " (3, array([ 0.96,  0.98,  0.91])),\n",
       " (4, array([ 0.96,  0.99,  0.9 ])),\n",
       " (5, array([ 0.96,  0.98,  0.92])),\n",
       " (6, array([ 0.96,  0.99,  0.91])),\n",
       " (7, array([ 0.96,  0.98,  0.92])),\n",
       " (8, array([ 0.96,  0.99,  0.91])),\n",
       " (9, array([ 0.97,  0.99,  0.93])),\n",
       " (10, array([ 0.97,  0.99,  0.93])),\n",
       " (20, array([ 0.95,  0.99,  0.89])),\n",
       " (40, array([ 0.95,  0.99,  0.87])),\n",
       " (50, array([ 0.95,  0.99,  0.87])),\n",
       " (100, array([ 0.93,  0.98,  0.83])),\n",
       " (150, array([ 0.91,  0.98,  0.78])),\n",
       " (200, array([ 0.9 ,  0.99,  0.74])),\n",
       " (250, array([ 0.86,  0.98,  0.63])),\n",
       " (300, array([ 0.82,  1.  ,  0.52]))]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ks = list(range(1,10+1)) + [20, 40, 50, 100, 150, 200, 250, 300]\n",
    "knn_res = k_nearest_neighbor(X, Y, ks)\n",
    "\n",
    "# means\n",
    "print(\"k, accuracy, precision, recall\")\n",
    "list((zip(ks, knn_res[:,:,0])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9, array([ 0.97,  0.99,  0.93]))"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find k-value with minimum distance from upper right corner\n",
    "dists = np.sqrt( (1-knn_res[:,1,0])**2 + (1-knn_res[:,2,0])**2 )\n",
    "ks[np.argmin(dists)], knn_res[np.argmin(dists), :, 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(1, array([ 0.01,  0.03,  0.03])),\n",
       " (2, array([ 0.02,  0.02,  0.06])),\n",
       " (3, array([ 0.02,  0.04,  0.05])),\n",
       " (4, array([ 0.02,  0.02,  0.06])),\n",
       " (5, array([ 0.03,  0.04,  0.07])),\n",
       " (6, array([ 0.03,  0.04,  0.07])),\n",
       " (7, array([ 0.02,  0.04,  0.05])),\n",
       " (8, array([ 0.02,  0.04,  0.06])),\n",
       " (9, array([ 0.02,  0.03,  0.05])),\n",
       " (10, array([ 0.02,  0.03,  0.05])),\n",
       " (20, array([ 0.02,  0.04,  0.06])),\n",
       " (40, array([ 0.02,  0.02,  0.05])),\n",
       " (50, array([ 0.02,  0.02,  0.05])),\n",
       " (100, array([ 0.02,  0.03,  0.06])),\n",
       " (150, array([ 0.03,  0.03,  0.06])),\n",
       " (200, array([ 0.02,  0.03,  0.05])),\n",
       " (250, array([ 0.02,  0.03,  0.05])),\n",
       " (300, array([ 0.05,  0.  ,  0.08]))]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# standard deviations\n",
    "list((zip(ks, knn_res[:,:,1])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# high precision value = there are no false positives. when algorithm says an instance is positive, it is\n",
    "# low recall value = not all positives are found.\n",
    "# we likely stop predicting positives all together at high k, because of imbalance in data set\n",
    "# k ~ N -> we just predict the majority class. which is negative ->> dont predict positives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaIAAAEUCAYAAABpmDIHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b7e8d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "knn_plot(knn_res, ks, 'knn')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# I checked that balancing does not really yield a difference\n",
    "# knn_res = k_nearest_neighbor(X_balanced, Y_balanced, ks)\n",
    "# knn_plot(knn_res, 'knn_bal')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "k, accuracy, precision, recall\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(1, array([ 0.94,  0.92,  0.91])),\n",
       " (2, array([ 0.93,  0.98,  0.83])),\n",
       " (3, array([ 0.95,  0.95,  0.91])),\n",
       " (4, array([ 0.94,  0.97,  0.87])),\n",
       " (5, array([ 0.94,  0.95,  0.89])),\n",
       " (6, array([ 0.94,  0.97,  0.86])),\n",
       " (7, array([ 0.95,  0.96,  0.89])),\n",
       " (8, array([ 0.94,  0.96,  0.86])),\n",
       " (9, array([ 0.94,  0.95,  0.89])),\n",
       " (10, array([ 0.94,  0.98,  0.87])),\n",
       " (20, array([ 0.94,  0.97,  0.85])),\n",
       " (40, array([ 0.93,  0.99,  0.82])),\n",
       " (50, array([ 0.93,  0.99,  0.81])),\n",
       " (100, array([ 0.89,  1.  ,  0.71])),\n",
       " (150, array([ 0.86,  1.  ,  0.61])),\n",
       " (200, array([ 0.8 ,  1.  ,  0.46]))]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# do NN in the first 6 dimensions of PCA space\n",
    "ks = list(range(1,10+1)) + [20, 40, 50, 100, 150, 200]\n",
    "knn_res = k_nearest_neighbor(Xp[:,:4], Y, ks)\n",
    "\n",
    "# means\n",
    "print(\"k, accuracy, precision, recall\")\n",
    "list((zip(ks, knn_res[:,:,0])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaIAAAEUCAYAAABpmDIHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWHkics45FytPRM4552L1/wEA/Kst80yQhgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116a1af98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "knn_plot(knn_res, ks, 'knn_pca')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import LinearSVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def lin_svm(X, Y, Cs):\n",
    "    # for each C, save [accuracy, precision, and recall] as mean and std\n",
    "    scores_svm = np.zeros((len(Cs), 3, 2))\n",
    "    for i, C in enumerate(Cs):\n",
    "        # instantiate lin SVM classifier\n",
    "        \n",
    "        svm_lin = LinearSVC(random_state=0)\n",
    "        svm_lin.set_params(C=C, loss='hinge')\n",
    "        svm_lin = make_pipeline(preprocessing.StandardScaler(), svm_lin)\n",
    "    \n",
    "        # standardise data to 0 mean and 1 variance\n",
    "        means_svm, stds_svm = cross_validate_with_statistics(svm_lin, X, Y)\n",
    "    \n",
    "        scores_svm[i, 0, 0] = means_svm['test_accuracy']\n",
    "        scores_svm[i, 1, 0] = means_svm['test_precision']\n",
    "        scores_svm[i, 2, 0] = means_svm['test_recall'] \n",
    "\n",
    "        scores_svm[i, 0, 1] = stds_svm['test_accuracy']\n",
    "        scores_svm[i, 1, 1] = stds_svm['test_precision']\n",
    "        scores_svm[i, 2, 1] = stds_svm['test_recall']\n",
    "    \n",
    "    return scores_svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# similar to knn_plot\n",
    "def svm_plot(knn_res, Cs, s):\n",
    "    \n",
    "    from matplotlib.ticker import FormatStrFormatter\n",
    "    \n",
    "    # precision recall plot\n",
    "    # recall\n",
    "    xpl = knn_res[:,2,0]\n",
    "    # precision\n",
    "    ypl = knn_res[:,1,0]\n",
    "\n",
    "    xerr = knn_res[:,2,1]\n",
    "    yerr = knn_res[:,1,1]\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.yaxis.set_major_formatter(FormatStrFormatter('%.2f'))\n",
    "\n",
    "    plt.errorbar(xpl, ypl, yerr, xerr, fmt='none', alpha=0.3, elinewidth=0.5)\n",
    "\n",
    "    plt.scatter(xpl, ypl, c=np.log10(Cs))\n",
    "    plt.colorbar(format=FormatStrFormatter('%i'), label='$\\log_{10} C$')\n",
    "\n",
    "    plt.plot([0,0.9999,1],[1,1,0], label='ideal', c='gray', alpha=0.8, linewidth=0.5)\n",
    "     \n",
    "    plt.annotate(r\"min\",(xpl[np.argmin(dists)]+0.004, ypl[np.argmin(dists)]), color='k')\n",
    "    plt.scatter(xpl[np.argmin(dists)], ypl[np.argmin(dists)], c='k', marker='x', alpha=0.5)\n",
    "\n",
    "    plt.ylabel('Precision')\n",
    "    plt.xlabel('Recall')\n",
    "    plt.ylim(0.88, 1.005)\n",
    "    plt.xlim(0.88, 1.01)\n",
    "    plt.tight_layout()\n",
    "    plt.savefig('figs/'+s+'.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "Cs = np.logspace(-5, 2, 30)\n",
    "svm_res = lin_svm(X, Y, Cs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0.00001, array([ 0.93148,  0.90466,  0.90268])),\n",
       " (0.00002, array([ 0.93148,  0.90466,  0.90268])),\n",
       " (0.00003, array([ 0.93148,  0.90466,  0.90268])),\n",
       " (0.00005, array([ 0.93324,  0.90554,  0.90856])),\n",
       " (0.00009, array([ 0.93324,  0.90554,  0.90856])),\n",
       " (0.00016, array([ 0.93499,  0.90921,  0.90856])),\n",
       " (0.00028, array([ 0.93675,  0.91319,  0.90856])),\n",
       " (0.00049, array([ 0.94377,  0.93252,  0.90856])),\n",
       " (0.00085, array([ 0.95254,  0.9516 ,  0.91318])),\n",
       " (0.00149, array([ 0.95777,  0.9635 ,  0.9188 ])),\n",
       " (0.00259, array([ 0.96654,  0.96819,  0.93994])),\n",
       " (0.00452, array([ 0.97359,  0.98856,  0.93994])),\n",
       " (0.00788, array([ 0.97534,  0.99444,  0.93994])),\n",
       " (0.01374, array([ 0.9807 ,  0.99444,  0.95458])),\n",
       " (0.02395, array([ 0.97895,  0.99444,  0.95073])),\n",
       " (0.04175, array([ 0.98246,  0.99444,  0.95958])),\n",
       " (0.07279, array([ 0.97892,  0.98468,  0.95958])),\n",
       " (0.12690, array([ 0.97719,  0.98413,  0.95558])),\n",
       " (0.22122, array([ 0.97365,  0.97093,  0.95943])),\n",
       " (0.38566, array([ 0.97365,  0.97111,  0.95972])),\n",
       " (0.67234, array([ 0.9719 ,  0.96614,  0.95972])),\n",
       " (1.17210, array([ 0.97014,  0.96026,  0.95972])),\n",
       " (2.04336, array([ 0.97187,  0.96493,  0.96019])),\n",
       " (3.56225, array([ 0.96836,  0.96464,  0.95094])),\n",
       " (6.21017, array([ 0.96836,  0.96881,  0.94506])),\n",
       " (10.82637, array([ 0.96836,  0.96895,  0.94506])),\n",
       " (18.87392, array([ 0.9719 ,  0.96969,  0.95759])),\n",
       " (32.90345, array([ 0.9719 ,  0.96907,  0.9562 ])),\n",
       " (57.36153, array([ 0.96313,  0.94167,  0.95787])),\n",
       " (100.00000, array([ 0.95786,  0.93883,  0.94847]))]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%precision 5\n",
    "list(zip(Cs, svm_res[:,:,0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.04175, array([ 0.98246,  0.99444,  0.95958]))"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# just like with k-NN, find min dist to upper right corner\n",
    "dists = np.sqrt( (1-svm_res[:,1,0])**2 + (1-svm_res[:,2,0])**2 )\n",
    "Cs[np.argmin(dists)], svm_res[np.argmin(dists), :, 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.04175, array([ 0.0111 ,  0.01667,  0.02918]))"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Cs[np.argmin(dists)], svm_res[np.argmin(dists), :, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaEAAAEUCAYAAACCr4kEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNQoFAgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11695b0f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_plot(svm_res, Cs, 'svm')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Kernel SVM (just for fun)\n",
    "\n",
    "probably wont do much good due to good results of linear SVM and noise in data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 10 folds for each of 42 candidates, totalling 420 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "[Parallel(n_jobs=1)]: Done 420 out of 420 | elapsed:   14.6s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=KFold(n_splits=10, random_state=42, shuffle=True),\n",
       "       error_score='raise',\n",
       "       estimator=SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False),\n",
       "       fit_params=None, iid=True, n_jobs=1,\n",
       "       param_grid={'C': [0.1, 1, 10, 100, 1000, 10000], 'gamma': [10, 1, 0.1, 0.01, 0.001, 0.0001, 1e-05]},\n",
       "       pre_dispatch='2*n_jobs', refit=False, return_train_score=True,\n",
       "       scoring=['accuracy', 'precision', 'recall'], verbose=1)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "from sklearn.model_selection import GridSearchCV, cross_val_score\n",
    "\n",
    "# quadratic exponential kernel\n",
    "svm = SVC(kernel=\"rbf\")\n",
    "\n",
    "p_grid = {\"C\": [0.1, 1, 10, 100, 1000, 10000],\n",
    "          \"gamma\": [10, 1, 0.1, 0.01, 0.001, 0.0001, 0.00001]}\n",
    "\n",
    "clf = GridSearchCV(estimator=svm, param_grid=p_grid, verbose=1, \n",
    "                   cv=KFold(10, shuffle=True, random_state=42),\n",
    "                  return_train_score=True, scoring=score_list,\n",
    "                  refit=False)\n",
    "clf.fit(X_scaled,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ignore warnings, bad results are expected for extreme parameter values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_grid_res = clf.cv_results_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max C, max gamma\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 3., -4.])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc = svm_grid_res[\"mean_test_accuracy\"]\n",
    "rec = svm_grid_res[\"mean_test_recall\"]\n",
    "prec = svm_grid_res[\"mean_test_precision\"]\n",
    "\n",
    "acc_err = svm_grid_res[\"std_test_accuracy\"]\n",
    "rec_err = svm_grid_res[\"std_test_recall\"]\n",
    "prec_err = svm_grid_res[\"std_test_precision\"]\n",
    "\n",
    "param_C = svm_grid_res[\"param_C\"]\n",
    "param_gamma = svm_grid_res[\"param_gamma\"]\n",
    "\n",
    "\n",
    "dists = np.sqrt( (1-rec)**2 + (1-prec)**2 )\n",
    "print(\"max C, max gamma\")\n",
    "np.log10([param_C[np.argmin(dists)], param_gamma[np.argmin(dists)]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.97540, 0.95550, 0.98146)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# minimal values\n",
    "acc[np.argmin(dists)], rec[np.argmin(dists)], prec[np.argmin(dists)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.01788, 0.03621, 0.02302)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_err[np.argmin(dists)], rec_err[np.argmin(dists)], prec_err[np.argmin(dists)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaEAAAEUCAYAAACCr4kEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nG5//uT8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11695bcf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_svm_grid(res, s):\n",
    "\n",
    "    import itertools\n",
    "    from matplotlib.ticker import FormatStrFormatter\n",
    "\n",
    "\n",
    "    xpl = res[\"mean_test_recall\"]\n",
    "    ypl = res[\"mean_test_precision\"]\n",
    "    \n",
    "    xerr = res[\"std_test_recall\"]\n",
    "    yerr = res[\"std_test_precision\"]\n",
    "    \n",
    "    \n",
    "    param_C = res[\"param_C\"]\n",
    "    param_gamma = res[\"param_gamma\"]\n",
    "    \n",
    "    markers = [\",\", \"o\", \"v\", \"^\", \"<\", \">\", \"1\", \n",
    "                \"s\", \"p\", \"*\", \"h\", \"+\", \"x\", \"d\"]\n",
    "\n",
    "    # choose marker by gamma and color by C\n",
    "    # note: param_C is masked array, this is why we have to convert it\n",
    "    \n",
    "    fig, ax = plt.subplots()\n",
    "    ax.yaxis.set_major_formatter(FormatStrFormatter('%.2f'))\n",
    "\n",
    "    plt.errorbar(xpl, ypl, yerr, xerr, fmt='none', alpha=0.3, elinewidth=0.2)\n",
    "    \n",
    "    \n",
    "    for i, g in enumerate(np.unique(param_gamma)):\n",
    "        plt.scatter(xpl[param_gamma==g], ypl[param_gamma==g], marker=markers[i], \n",
    "                    c=np.log10(np.array(param_C[param_gamma==g]).astype('float')), label=np.log10(g))\n",
    "    \n",
    "    \n",
    "    \n",
    "    plt.colorbar(format=FormatStrFormatter('%i'), label=r'$\\log_{10} C$')\n",
    "    \n",
    "    plt.annotate(r\"min\",(xpl[np.argmin(dists)]+0.004, ypl[np.argmin(dists)]), color='k')\n",
    "    plt.scatter(xpl[np.argmin(dists)], ypl[np.argmin(dists)], c='k', marker='x', alpha=0.3)\n",
    "\n",
    "    plt.plot([0,0.9999,1],[1,1,0], c='gray', alpha=0.8, linewidth=0.5)\n",
    "    \n",
    "    \n",
    "    plt.ylabel('Precision')\n",
    "    plt.xlabel('Recall')\n",
    "    plt.ylim(0.8, 1.02)\n",
    "    plt.xlim(0.88, 1.01)\n",
    "    plt.legend(loc='lower left', fontsize=\"small\", title=r'$\\log_{10} \\gamma$')\n",
    "    plt.tight_layout()\n",
    "    plt.savefig('figs/'+s+'.pdf')\n",
    "\n",
    "plot_svm_grid(svm_grid_res, 'svm_grid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average difference of 0.002279 with std. dev. of 0.003800.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcMAAAEoCAYAAAAkBZqfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxhhj2j2rDI0xxrR7/x+/hPQYWj9nZgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b551358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# SIDENOTE:\n",
    "# cross validation above is non-nested, when nested cv is preferred\n",
    "# have checked below that extra model optimism does not matter\n",
    "# in case of kernel SVM\n",
    "# therefore it was okay that we neglected this above\n",
    "# and below for computational concerns\n",
    "# noise so so large that the extra model optimism probably does not matter \n",
    "# vary parameters just for trends in precision recall curves\n",
    "# and not for finding ideal parameters\n",
    "# code from:\n",
    "# http://scikit-learn.org/stable/auto_examples/model_selection/plot_nested_cross_validation_iris.html#sphx-glr-auto-examples-model-selection-plot-nested-cross-validation-iris-py\n",
    "\n",
    "# Number of random trials\n",
    "NUM_TRIALS = 30\n",
    "\n",
    "# Set up possible values of parameters to optimize over\n",
    "p_grid = {\"C\": [1, 10, 100],\n",
    "          \"gamma\": [.01, .1]}\n",
    "\n",
    "# We will use a Support Vector Classifier with \"rbf\" kernel\n",
    "svm = SVC(kernel=\"rbf\")\n",
    "\n",
    "# Arrays to store scores\n",
    "non_nested_scores = np.zeros(NUM_TRIALS)\n",
    "nested_scores = np.zeros(NUM_TRIALS)\n",
    "\n",
    "# Loop for each trial\n",
    "for i in range(NUM_TRIALS):\n",
    "\n",
    "    # Choose cross-validation techniques for the inner and outer loops,\n",
    "    # independently of the dataset.\n",
    "    # E.g \"LabelKFold\", \"LeaveOneOut\", \"LeaveOneLabelOut\", etc.\n",
    "    inner_cv = KFold(n_splits=4, shuffle=True, random_state=i)\n",
    "    outer_cv = KFold(n_splits=4, shuffle=True, random_state=i)\n",
    "\n",
    "    # Non_nested parameter search and scoring\n",
    "    clf = GridSearchCV(estimator=svm, param_grid=p_grid, cv=inner_cv)\n",
    "    clf.fit(X_scaled, Y)\n",
    "    non_nested_scores[i] = clf.best_score_\n",
    "\n",
    "    # Nested CV with parameter optimization\n",
    "    nested_score = cross_val_score(clf, X=X_scaled, y=Y, cv=outer_cv)\n",
    "    nested_scores[i] = nested_score.mean()\n",
    "\n",
    "score_difference = non_nested_scores - nested_scores\n",
    "\n",
    "print(\"Average difference of {0:6f} with std. dev. of {1:6f}.\"\n",
    "      .format(score_difference.mean(), score_difference.std()))\n",
    "\n",
    "# Plot scores on each trial for nested and non-nested CV\n",
    "plt.figure()\n",
    "plt.subplot(211)\n",
    "non_nested_scores_line, = plt.plot(non_nested_scores, color='r')\n",
    "nested_line, = plt.plot(nested_scores, color='b')\n",
    "plt.ylabel(\"score\", fontsize=\"14\")\n",
    "plt.legend([non_nested_scores_line, nested_line],\n",
    "           [\"Non-Nested CV\", \"Nested CV\"],\n",
    "           bbox_to_anchor=(0, .4, .5, 0))\n",
    "plt.title(\"Non-Nested and Nested Cross Validation on Iris Dataset\",\n",
    "          x=.5, y=1.1, fontsize=\"15\")\n",
    "\n",
    "# Plot bar chart of the difference.\n",
    "plt.subplot(212)\n",
    "difference_plot = plt.bar(range(NUM_TRIALS), score_difference)\n",
    "plt.xlabel(\"Individual Trial #\")\n",
    "plt.legend([difference_plot],\n",
    "           [\"Non-Nested CV - Nested CV Score\"],\n",
    "           bbox_to_anchor=(0, 1, .8, 0))\n",
    "plt.ylabel(\"score difference\", fontsize=\"14\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple ANN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jannikkossen/miniconda3/envs/foml17/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n"
     ]
    }
   ],
   "source": [
    "# adapted from http://adventuresinmachinelearning.com/python-tensorflow-tutorial/\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(\n",
    "            X_scaled, Y, test_size=0.33,\n",
    "            shuffle=True, random_state=42)\n",
    "\n",
    "# Python optimisation variables\n",
    "learning_rate = 0.05\n",
    "# how often do we pass all training data\n",
    "epochs = 200\n",
    "# how big is a batch (backprop after each batch)\n",
    "batch_size = 50\n",
    "\n",
    "# declare training data placeholders\n",
    "# 30 input dimensions, binary outputs\n",
    "x = tf.placeholder(tf.float32, [None, 30], name='x')\n",
    "y = tf.placeholder(tf.float32, [None, 2], name='y')\n",
    "# None: unspecified number of input samples\n",
    "\n",
    "# setup weights and biases for 3 layer network\n",
    "W1 = tf.Variable(tf.random_normal([30, 300], stddev=0.03), name='W1')\n",
    "b1 = tf.Variable(tf.random_normal([300]), name='b1')\n",
    "W2 = tf.Variable(tf.random_normal([300, 2], stddev=0.03), name='W2')\n",
    "b2 = tf.Variable(tf.random_normal([2]), name='b2')\n",
    "\n",
    "# setup node inputs and activation functions of hidden layer\n",
    "hidden_out = tf.add(tf.matmul(x, W1), b1)\n",
    "hidden_out = tf.nn.relu(hidden_out)\n",
    "\n",
    "# calculate the output layer, use softmax\n",
    "y_ = tf.nn.softmax(tf.add(tf.matmul(hidden_out, W2), b2))\n",
    "\n",
    "# calculate cross-entropy loss\n",
    "y_clipped = tf.clip_by_value(y_, 1e-10, 0.9999999)\n",
    "\n",
    "cross_entropy = -tf.reduce_mean(tf.reduce_sum(y * tf.log(y_clipped) \n",
    "                                              + (1 - y) * tf.log(1 - y_clipped), axis=1))\n",
    "# sum over the first axis in reduce_sum\n",
    "# 0th axis = axis of the samples\n",
    "# 1st axis = number of nodes\n",
    "# so to get cross entropy we sum over all output nodes\n",
    "\n",
    "# afterwards we have a 1D-tensor (i.e. loss over all nodes (scalar) for each input sample) \n",
    "# -> mean of that is our cross-entropy loss\n",
    "\n",
    "# set up optimiser\n",
    "optimiser = tf.train.GradientDescentOptimizer(learning_rate=learning_rate).minimize(cross_entropy)\n",
    "\n",
    "# define accuracy assesment operation\n",
    "correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))\n",
    "# argmax(y_,1) is argmax of the output nodes, i.e. number between 0-1, our prediction\n",
    "# & since y is one-hot vector, we also have to do argmax here\n",
    "# second argmument gives axis. remember: first axis is number of samples\n",
    "\n",
    "\n",
    "# define metrics\n",
    "_, accuracy = tf.metrics.accuracy(tf.cast(tf.argmax(y, 1), tf.float32),\n",
    "                                 tf.cast(tf.argmax(y_,1), tf.float32))\n",
    "\n",
    "_, precision = tf.metrics.precision(tf.cast(tf.argmax(y, 1), tf.float32),\n",
    "                                 tf.cast(tf.argmax(y_,1), tf.float32))\n",
    "\n",
    "_, recall = tf.metrics.recall(tf.cast(tf.argmax(y, 1), tf.float32),\n",
    "                                 tf.cast(tf.argmax(y_,1), tf.float32))\n",
    "\n",
    "# set up initialiser operation\n",
    "\n",
    "init_op = tf.global_variables_initializer()\n",
    "init_l = tf.local_variables_initializer()\n",
    "\n",
    "\n",
    "# gives proportion of correct decisions\n",
    "# cast boolean to tf.float32, otherwise tensorflow cries\n",
    "\n",
    "# add a summary to store the accuracy\n",
    "tf.summary.scalar('accuracy', accuracy)\n",
    "tf.summary.scalar('precision', precision)\n",
    "tf.summary.scalar('recall', recall)\n",
    "\n",
    "merged = tf.summary.merge_all()\n",
    "writer = tf.summary.FileWriter('/')\n",
    "\n",
    "# Y labels to one hot encoding\n",
    "def to_one_hot(Y):\n",
    "    return np.stack([1-Y, Y], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0, cost = 2.774\n",
      "Epoch: 1, cost = 1.124\n",
      "Epoch: 2, cost = 0.292\n",
      "Epoch: 3, cost = 0.234\n",
      "Epoch: 4, cost = 0.220\n",
      "Epoch: 5, cost = 0.219\n",
      "Epoch: 6, cost = 0.196\n",
      "Epoch: 7, cost = 0.208\n",
      "Epoch: 8, cost = 0.183\n",
      "Epoch: 9, cost = 0.197\n",
      "Epoch: 10, cost = 0.168\n",
      "Epoch: 11, cost = 0.167\n",
      "Epoch: 12, cost = 0.184\n",
      "Epoch: 13, cost = 0.156\n",
      "Epoch: 14, cost = 0.146\n",
      "Epoch: 15, cost = 0.148\n",
      "Epoch: 16, cost = 0.149\n",
      "Epoch: 17, cost = 0.141\n",
      "Epoch: 18, cost = 0.152\n",
      "Epoch: 19, cost = 0.144\n",
      "Epoch: 20, cost = 0.140\n",
      "Epoch: 21, cost = 0.152\n",
      "Epoch: 22, cost = 0.141\n",
      "Epoch: 23, cost = 0.145\n",
      "Epoch: 24, cost = 0.136\n",
      "Epoch: 25, cost = 0.126\n",
      "Epoch: 26, cost = 0.142\n",
      "Epoch: 27, cost = 0.134\n",
      "Epoch: 28, cost = 0.134\n",
      "Epoch: 29, cost = 0.132\n",
      "Epoch: 30, cost = 0.081\n",
      "Epoch: 31, cost = 0.137\n",
      "Epoch: 32, cost = 0.117\n",
      "Epoch: 33, cost = 0.128\n",
      "Epoch: 34, cost = 0.127\n",
      "Epoch: 35, cost = 0.114\n",
      "Epoch: 36, cost = 0.126\n",
      "Epoch: 37, cost = 0.123\n",
      "Epoch: 38, cost = 0.122\n",
      "Epoch: 39, cost = 0.115\n",
      "Epoch: 40, cost = 0.119\n",
      "Epoch: 41, cost = 0.121\n",
      "Epoch: 42, cost = 0.111\n",
      "Epoch: 43, cost = 0.122\n",
      "Epoch: 44, cost = 0.119\n",
      "Epoch: 45, cost = 0.099\n",
      "Epoch: 46, cost = 0.116\n",
      "Epoch: 47, cost = 0.114\n",
      "Epoch: 48, cost = 0.119\n",
      "Epoch: 49, cost = 0.114\n",
      "Epoch: 50, cost = 0.116\n",
      "Epoch: 51, cost = 0.125\n",
      "Epoch: 52, cost = 0.115\n",
      "Epoch: 53, cost = 0.118\n",
      "Epoch: 54, cost = 0.111\n",
      "Epoch: 55, cost = 0.107\n",
      "Epoch: 56, cost = 0.117\n",
      "Epoch: 57, cost = 0.109\n",
      "Epoch: 58, cost = 0.144\n",
      "Epoch: 59, cost = 0.103\n",
      "Epoch: 60, cost = 0.098\n",
      "Epoch: 61, cost = 0.107\n",
      "Epoch: 62, cost = 0.109\n",
      "Epoch: 63, cost = 0.105\n",
      "Epoch: 64, cost = 0.133\n",
      "Epoch: 65, cost = 0.119\n",
      "Epoch: 66, cost = 0.105\n",
      "Epoch: 67, cost = 0.113\n",
      "Epoch: 68, cost = 0.106\n",
      "Epoch: 69, cost = 0.103\n",
      "Epoch: 70, cost = 0.059\n",
      "Epoch: 71, cost = 0.106\n",
      "Epoch: 72, cost = 0.101\n",
      "Epoch: 73, cost = 0.104\n",
      "Epoch: 74, cost = 0.101\n",
      "Epoch: 75, cost = 0.106\n",
      "Epoch: 76, cost = 0.110\n",
      "Epoch: 77, cost = 0.076\n",
      "Epoch: 78, cost = 0.102\n",
      "Epoch: 79, cost = 0.099\n",
      "Epoch: 80, cost = 0.071\n",
      "Epoch: 81, cost = 0.063\n",
      "Epoch: 82, cost = 0.107\n",
      "Epoch: 83, cost = 0.107\n",
      "Epoch: 84, cost = 0.079\n",
      "Epoch: 85, cost = 0.100\n",
      "Epoch: 86, cost = 0.094\n",
      "Epoch: 87, cost = 0.095\n",
      "Epoch: 88, cost = 0.081\n",
      "Epoch: 89, cost = 0.096\n",
      "Epoch: 90, cost = 0.084\n",
      "Epoch: 91, cost = 0.094\n",
      "Epoch: 92, cost = 0.083\n",
      "Epoch: 93, cost = 0.093\n",
      "Epoch: 94, cost = 0.107\n",
      "Epoch: 95, cost = 0.093\n",
      "Epoch: 96, cost = 0.099\n",
      "Epoch: 97, cost = 0.086\n",
      "Epoch: 98, cost = 0.097\n",
      "Epoch: 99, cost = 0.086\n",
      "Epoch: 100, cost = 0.119\n",
      "Epoch: 101, cost = 0.094\n",
      "Epoch: 102, cost = 0.087\n",
      "Epoch: 103, cost = 0.104\n",
      "Epoch: 104, cost = 0.089\n",
      "Epoch: 105, cost = 0.091\n",
      "Epoch: 106, cost = 0.086\n",
      "Epoch: 107, cost = 0.087\n",
      "Epoch: 108, cost = 0.077\n",
      "Epoch: 109, cost = 0.086\n",
      "Epoch: 110, cost = 0.081\n",
      "Epoch: 111, cost = 0.087\n",
      "Epoch: 112, cost = 0.091\n",
      "Epoch: 113, cost = 0.082\n",
      "Epoch: 114, cost = 0.088\n",
      "Epoch: 115, cost = 0.109\n",
      "Epoch: 116, cost = 0.077\n",
      "Epoch: 117, cost = 0.087\n",
      "Epoch: 118, cost = 0.087\n",
      "Epoch: 119, cost = 0.081\n",
      "Epoch: 120, cost = 0.087\n",
      "Epoch: 121, cost = 0.083\n",
      "Epoch: 122, cost = 0.079\n",
      "Epoch: 123, cost = 0.075\n",
      "Epoch: 124, cost = 0.083\n",
      "Epoch: 125, cost = 0.059\n",
      "Epoch: 126, cost = 0.082\n",
      "Epoch: 127, cost = 0.085\n",
      "Epoch: 128, cost = 0.081\n",
      "Epoch: 129, cost = 0.079\n",
      "Epoch: 130, cost = 0.083\n",
      "Epoch: 131, cost = 0.080\n",
      "Epoch: 132, cost = 0.070\n",
      "Epoch: 133, cost = 0.078\n",
      "Epoch: 134, cost = 0.077\n",
      "Epoch: 135, cost = 0.074\n",
      "Epoch: 136, cost = 0.072\n",
      "Epoch: 137, cost = 0.077\n",
      "Epoch: 138, cost = 0.081\n",
      "Epoch: 139, cost = 0.067\n",
      "Epoch: 140, cost = 0.078\n",
      "Epoch: 141, cost = 0.049\n",
      "Epoch: 142, cost = 0.068\n",
      "Epoch: 143, cost = 0.071\n",
      "Epoch: 144, cost = 0.073\n",
      "Epoch: 145, cost = 0.068\n",
      "Epoch: 146, cost = 0.077\n",
      "Epoch: 147, cost = 0.070\n",
      "Epoch: 148, cost = 0.074\n",
      "Epoch: 149, cost = 0.056\n",
      "Epoch: 150, cost = 0.072\n",
      "Epoch: 151, cost = 0.074\n",
      "Epoch: 152, cost = 0.069\n",
      "Epoch: 153, cost = 0.071\n",
      "Epoch: 154, cost = 0.072\n",
      "Epoch: 155, cost = 0.068\n",
      "Epoch: 156, cost = 0.074\n",
      "Epoch: 157, cost = 0.061\n",
      "Epoch: 158, cost = 0.069\n",
      "Epoch: 159, cost = 0.066\n",
      "Epoch: 160, cost = 0.073\n",
      "Epoch: 161, cost = 0.065\n",
      "Epoch: 162, cost = 0.066\n",
      "Epoch: 163, cost = 0.054\n",
      "Epoch: 164, cost = 0.068\n",
      "Epoch: 165, cost = 0.065\n",
      "Epoch: 166, cost = 0.062\n",
      "Epoch: 167, cost = 0.077\n",
      "Epoch: 168, cost = 0.069\n",
      "Epoch: 169, cost = 0.063\n",
      "Epoch: 170, cost = 0.066\n",
      "Epoch: 171, cost = 0.058\n",
      "Epoch: 172, cost = 0.062\n",
      "Epoch: 173, cost = 0.063\n",
      "Epoch: 174, cost = 0.063\n",
      "Epoch: 175, cost = 0.064\n",
      "Epoch: 176, cost = 0.057\n",
      "Epoch: 177, cost = 0.068\n",
      "Epoch: 178, cost = 0.061\n",
      "Epoch: 179, cost = 0.056\n",
      "Epoch: 180, cost = 0.047\n",
      "Epoch: 181, cost = 0.053\n",
      "Epoch: 182, cost = 0.061\n",
      "Epoch: 183, cost = 0.047\n",
      "Epoch: 184, cost = 0.058\n",
      "Epoch: 185, cost = 0.061\n",
      "Epoch: 186, cost = 0.052\n",
      "Epoch: 187, cost = 0.057\n",
      "Epoch: 188, cost = 0.055\n",
      "Epoch: 189, cost = 0.053\n",
      "Epoch: 190, cost = 0.050\n",
      "Epoch: 191, cost = 0.055\n",
      "Epoch: 192, cost = 0.061\n",
      "Epoch: 193, cost = 0.054\n",
      "Epoch: 194, cost = 0.057\n",
      "Epoch: 195, cost = 0.041\n",
      "Epoch: 196, cost = 0.049\n",
      "Epoch: 197, cost = 0.053\n",
      "Epoch: 198, cost = 0.051\n",
      "Epoch: 199, cost = 0.046\n",
      "\n",
      "Training complete!\n",
      "Accuracy 0.980\n",
      "Precision 0.970\n",
      "Recall 0.975\n"
     ]
    }
   ],
   "source": [
    "# set up training & testing\n",
    "with tf.Session() as sess:\n",
    "    sess.run(init_op)\n",
    "    sess.run(init_l)\n",
    "    #t training\n",
    "    # how many mini-batches for one epoch\n",
    "    total_batch = int(len(Y_train) / batch_size) \n",
    "    for epoch in range(epochs):\n",
    "        avg_cost = 0\n",
    "        for i in range(total_batch):\n",
    "            batch_x = X_train[i*batch_size:1+(i+1)*batch_size]\n",
    "            batch_y = Y_train[i*batch_size:1+(i+1)*batch_size]\n",
    "            # convert to one-hot\n",
    "\n",
    "            batch_y = to_one_hot(batch_y)\n",
    "            _, c = sess.run([optimiser, cross_entropy],\n",
    "                           feed_dict={x:batch_x, y:batch_y})\n",
    "            # both operations optimiser and cross_entropy will be executed\n",
    "            # throw away output from optimiser operation, we dont really care about that\n",
    "            # but we do care about output from cross entropy\n",
    "            # code also works without explicitly giving cross_entropy, \n",
    "            # since optimiser needs it, but if we want output from cross_entropy, \n",
    "            # we have to specify it as an additional task\n",
    "            avg_cost += c/total_batch\n",
    "        \n",
    "        print(\"Epoch: {}, cost = {:.3f}\".format(epoch, avg_cost))\n",
    "        summary = sess.run(merged, feed_dict={x: X_test, y: to_one_hot(Y_test)})\n",
    "        writer.add_summary(summary, epoch)\n",
    "        \n",
    "        # shuffle data after each epoch (similar to how next_batch() of mnist would)\n",
    "        perm = np.arange(len(Y_train))\n",
    "        np.random.shuffle(perm)\n",
    "        X_train = X_train[perm]\n",
    "        Y_train = Y_train[perm]\n",
    "    \n",
    "    print(\"\\nTraining complete!\")\n",
    "    writer.add_graph(sess.graph)\n",
    "\n",
    "    # testing    \n",
    "    test_accuracy, test_precision, test_recall = sess.run([accuracy, precision, recall] , \n",
    "                                                         feed_dict={x: X_test, y: to_one_hot(Y_test)})\n",
    "    print(\"Accuracy {:.3f}\". format(test_accuracy))\n",
    "    print(\"Precision {:.3f}\". format(test_precision))\n",
    "    print(\"Recall {:.3f}\". format(test_recall))"
   ]
  }
 ],
 "metadata": {
  "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}