Will Couples Stay Together.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Will Couples Stay Together?\n",
    "\n",
    "Because I love asking the important questions. Also explain your predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/home/stephen/miniconda3/lib/python3.6/importlib/_bootstrap.py:205: 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": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as graph\n",
    "from IPython.display import display, Markdown\n",
    "\n",
    "from DataScienceUtils import GetSubsetTransform\n",
    "from DataScienceUtils.models import PreTrainedVoteEnsemble\n",
    "from DataScienceUtils.graphing import plot_roc_curve\n",
    "from lime.lime_tabular import LimeTabularExplainer\n",
    "\n",
    "from sklearn.base import ClassifierMixin, TransformerMixin\n",
    "from sklearn.ensemble import RandomForestClassifier, VotingClassifier\n",
    "from sklearn.metrics import f1_score, matthews_corrcoef, accuracy_score, roc_auc_score\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV, RepeatedStratifiedKFold\n",
    "from sklearn.pipeline import make_pipeline, Pipeline, FeatureUnion\n",
    "from sklearn.preprocessing import LabelEncoder, StandardScaler, MinMaxScaler, OneHotEncoder\n",
    "\n",
    "import keras\n",
    "from keras.models import Model, Sequential\n",
    "from keras.layers import Dense, Input, Dropout\n",
    "from keras.optimizers import Adam\n",
    "from keras.callbacks import EarlyStopping\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "\n",
    "graph.style.use('fivethirtyeight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1473, 22)\n"
     ]
    },
    {
     "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>caseid_new</th>\n",
       "      <th>marital_status</th>\n",
       "      <th>age</th>\n",
       "      <th>partner_age</th>\n",
       "      <th>education</th>\n",
       "      <th>partner_education</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>children</th>\n",
       "      <th>visits_relatives</th>\n",
       "      <th>gender</th>\n",
       "      <th>...</th>\n",
       "      <th>msa</th>\n",
       "      <th>rent</th>\n",
       "      <th>political</th>\n",
       "      <th>religion</th>\n",
       "      <th>work</th>\n",
       "      <th>lgbtq</th>\n",
       "      <th>success</th>\n",
       "      <th>age_diff_abs</th>\n",
       "      <th>gender_older</th>\n",
       "      <th>education_difference</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>22526</td>\n",
       "      <td>living with partner</td>\n",
       "      <td>52</td>\n",
       "      <td>48</td>\n",
       "      <td>bachelor's degree or higher</td>\n",
       "      <td>some college</td>\n",
       "      <td>hispanic</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>female</td>\n",
       "      <td>...</td>\n",
       "      <td>metro</td>\n",
       "      <td>rented for cash</td>\n",
       "      <td>democrat</td>\n",
       "      <td>catholic</td>\n",
       "      <td>working - as a paid employee</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23286</td>\n",
       "      <td>living with partner</td>\n",
       "      <td>28</td>\n",
       "      <td>30</td>\n",
       "      <td>bachelor's degree or higher</td>\n",
       "      <td>bachelor's degree or higher</td>\n",
       "      <td>white, non-hispanic</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>female</td>\n",
       "      <td>...</td>\n",
       "      <td>metro</td>\n",
       "      <td>rented for cash</td>\n",
       "      <td>democrat</td>\n",
       "      <td>jewish</td>\n",
       "      <td>working - as a paid employee</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>26315</td>\n",
       "      <td>never married</td>\n",
       "      <td>31</td>\n",
       "      <td>40</td>\n",
       "      <td>some college</td>\n",
       "      <td>high school</td>\n",
       "      <td>white, non-hispanic</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>male</td>\n",
       "      <td>...</td>\n",
       "      <td>metro</td>\n",
       "      <td>owned or being bought by you or someone in you...</td>\n",
       "      <td>democrat</td>\n",
       "      <td>other non-christian, please specify:</td>\n",
       "      <td>working - as a paid employee</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   caseid_new       marital_status  age  partner_age  \\\n",
       "0       22526  living with partner   52           48   \n",
       "1       23286  living with partner   28           30   \n",
       "2       26315        never married   31           40   \n",
       "\n",
       "                     education            partner_education  \\\n",
       "0  bachelor's degree or higher                 some college   \n",
       "1  bachelor's degree or higher  bachelor's degree or higher   \n",
       "2                 some college                  high school   \n",
       "\n",
       "             ethnicity  children  visits_relatives  gender  \\\n",
       "0             hispanic       0.0                 0  female   \n",
       "1  white, non-hispanic       0.0                 0  female   \n",
       "2  white, non-hispanic       0.0                 1    male   \n",
       "\n",
       "           ...             msa  \\\n",
       "0          ...           metro   \n",
       "1          ...           metro   \n",
       "2          ...           metro   \n",
       "\n",
       "                                                rent political  \\\n",
       "0                                    rented for cash  democrat   \n",
       "1                                    rented for cash  democrat   \n",
       "2  owned or being bought by you or someone in you...  democrat   \n",
       "\n",
       "                               religion                          work lgbtq  \\\n",
       "0                              catholic  working - as a paid employee     0   \n",
       "1                                jewish  working - as a paid employee     0   \n",
       "2  other non-christian, please specify:  working - as a paid employee     0   \n",
       "\n",
       "  success  age_diff_abs  gender_older  education_difference  \n",
       "0     1.0             4             1                     1  \n",
       "1     1.0             2             0                     0  \n",
       "2     0.0             9             1                     1  \n",
       "\n",
       "[3 rows x 22 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1473 entries, 0 to 1472\n",
      "Data columns (total 22 columns):\n",
      "caseid_new              1473 non-null int64\n",
      "marital_status          1473 non-null object\n",
      "age                     1473 non-null int64\n",
      "partner_age             1473 non-null int64\n",
      "education               1473 non-null object\n",
      "partner_education       1473 non-null object\n",
      "ethnicity               1473 non-null object\n",
      "children                1473 non-null float64\n",
      "visits_relatives        1473 non-null int64\n",
      "gender                  1473 non-null object\n",
      "house                   1473 non-null object\n",
      "income                  1473 non-null object\n",
      "msa                     1473 non-null object\n",
      "rent                    1473 non-null object\n",
      "political               1473 non-null object\n",
      "religion                1473 non-null object\n",
      "work                    1473 non-null object\n",
      "lgbtq                   1473 non-null int64\n",
      "success                 1473 non-null float64\n",
      "age_diff_abs            1473 non-null int64\n",
      "gender_older            1473 non-null int64\n",
      "education_difference    1473 non-null int64\n",
      "dtypes: float64(2), int64(8), object(12)\n",
      "memory usage: 253.2+ KB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Load data (Came from Stanford https://data.stanford.edu/hcmst)\n",
    "df = pd.read_csv('data/couples.csv')\n",
    "\n",
    "print(df.shape)\n",
    "display(df.head(3))\n",
    "display(df.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "## String Features"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693e23eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693de4550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCAAAADgCAYAAADSSxfCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3Xm4JFV9//H3BwZBQAdBZBcQxgVjgmhcEmNQ3GOCMYAoCgLRENDELQYjUaKCS35xISgQ2V3AFUEkKqIoGlkUCKs4IwyZYVdkHRZHvr8/6lynud51uN13mffrefrpqlOnqr5V91Z19bdPnUpVIUmSJEmS1E+rTXcAkiRJkiRp7jMBIUmSJEmS+s4EhCRJkiRJ6jsTEJIkSZIkqe9MQEiSJEmSpL4zASFJkiRJkvrOBIQkSZrRkmyVpJI8Z7pjAUhydpKjpzsOSZJmGxMQkiTNAEmOb1+yK8nyJNcmOTLJBlOw7IOSLJ6CMB9qHDv2bOPw1y7THd9wY+y3VwJvG3A4kiTNevOmOwBJkvQ75wC70X0+Pw04GtgC+IuVWViSMA2f9UkeVlX3j1FlB+CGYWW/7mNIU6qqbp3uGCRJmo1sASFJ0sxxf1XdWFVLq+pU4OPAS5I8HCDJIUmuTLIsyZLWQmL+0MxJXt9aTzwvyUXAfcDfAu8HtuxpbXBwq784yfuSfCLJrUluSvKxJA9KWiR5c5KfJbk3ycIk7+6t05bzgSSfSvIrukTKWG5p29n7uq9nebslWdTW9z/AHw6LZ6glxebDypcneX3P+GOSHNe2694kVyXZp01Lkk8n+UWSe5JcneTQJGsO7csx9tuDbsFIskaSDyW5Lsn9Sa5I8pphsVWS/ZN8JsmdSZYmedc4+0mSpDnFFhCSJM1c99D9WDCvZ/yNwBJgG+CTwGHAXj3zrAZ8mO4WgWuB3wBbAnsAf9zq3NVT/82t/jOBpwKfAy4DjgFoX7r3Bt4CXAw8CTgSWAv4157l/APwUeDZPITriyRPBU4CPgIcDzwZ+MRKLOfhwPfp9tkewNXAtsD6Q1WAm4HXADfRJTmOottf7wW+ADyR0fdbr0OBfYD9gP8FdgE+m+Smqjqrp957gYOAg4GXAIcnOX9YHUmS5iwTEJIkzUBJtgMOAM6rqjsBquoDPVUWt1/QT06yd1U9MDQr8PaqOqdnWXcBv62qG0dY1TlV9aE2vDDJ3sALgGOSrA28E3hlVX2z1bkmyUF0iY/eBMQFVXXwBDfvqiQ1rOzxVXU98Hbg3Kp6V0/dTYH/nOCyh7wG2BrYtqqWtrKrhya2/fXunvqLk2wD7A+8t6ruGWe/AdD20T8Ab62qL7XiQ5P8cVt+b3LhC1X16Tb8ySRvotvXJiAkSasEExCSJM0cO7YvvasDa9J9Mf27oYlJXknXEmFb4JF0rR0eBmwMXN+znAsmsc6Lh41fT/fFHbrWBw8HvjIsYbA6sFaSDavqllZ2/iTW+WJg+Jf6m9r7dvz+F/IfTmLZQ54GXNGTfPg9Sd5Ad4vKVsA6dNdFk709dVu6v8EPhpV/Hxh+i8VI+3qjSa5PkqRZywSEJEkzx3l0t1MsB67v7cgxyTOBLwEfBP6JrtPGZwEn0H0BHvLbqrp3Eusc3llkseJL+ND7rsDPR5i3tzPGuyexzsVjJQYmoLe1RzeQrM4kkgdJdqW7heVAumTBHXTbechDiGs8Y+1rSZLmPBMQkiTNHPdU1aJRpj0H+GVVHTRUMIlHV95P12phsi4H7gUeV1VnrMT8K+MK4E+Glf3psPGb2/umdP1hAGxPT0IC+CmwT5LNR0l2PBe4qKo+OlSQZKthdSay3xbRdfb5XLq+M4b8+bBxSZJWeSYgJEmaHa4CNkyyL/A9uoTE/hOc9xpg4yTPBhYCy6pq2XgzVdVdSQ6l69OggO/QXTs8BXhqVf3zSmwHdNuxfFjZHS2mjwEXJDmErnXHk+n6hei1iK6DzYOTvBV4NF1HkL23iZxE13/FaUneCfwCeBzw6Kr6At3+3DfJznSJgpcDrxy2nnH3W1UtS3IY8P4kt7CiE8qdgRdOZqdIkjTX2exPkqRZoKpOp7s94FDgUmB3ulsxJuJrdLdvfAO4he6L+UTX+366J2q8ge7L9Q+BtwKLJ7qMEVwI3DDstX9b30/pOpDcnW47D2zr641pOfAq4DHARXS3UrybFbdm0BIFQ60QTgaubPUe3qocBXwGOK4t45l0T6foNdH99m7g03SPTb0MeC3wWp9uIUnSg6VqeCfUkiRJkiRJU8sWEJIkSZIkqe8GloBIsl6SLyf5WZIrkzw7yfpJzkyysL0/qtVNksOSLEpySZIdBhWnJEmSJEmaeoNsAfEJ4JtV9UTgj+juxTwQOKuqFtA98/vAVvelwIL2eiNwxADjlCRJkiRJU2wgfUAkmQ9cTPcYr+opvwrYsapuSLIJcHZVPSHJUW34pOH1hua9/fbb7bxCkiRJkqQZaP78+RleNqgWEFvT9R59XJKLkhydZB1go56kwo3ARm14M1Y81xtgaSuTJEmSJEmz0KASEPOAHYAjquqpwN2suN0CgNYywlYNkiRJkiTNQYNKQCwFllbVeW38y3QJiZvarRe095vb9OuALXrm37yVSZIkSZKkWWggCYiquhFYkuQJrWgn4ArgNGCvVrYXcGobPg3Ysz0N41nA7b39P0iSJEmSpNll3gDX9Wbgc0keBlwN7E2XAPlikn2Ba4HdWt0zgJcBi4Blra4kSZIkSZqlBvIUjH7wKRiSJEmSJM1M0/kUDEmSJEmStAozASFJkiRJkvpukH1ASNKMdOdbXz3dIUgz2iM+dtJ0hyBJkuYAW0BIkiRJkqS+MwEhSZIkSZL6zgSEJEmSJEnqOxMQkiRJkiSp70xASJIkSZKkvjMBIUmSJEmS+s4EhCRJkiRJ6jsTEJIkSZIkqe9MQEiSJEmSpL4zASFJkiRJkvrOBIQkSZIkSeo7ExCSJEmSJKnvTEBIkiRJkqS+MwEhSZIkSZL6zgSEJEmSJEnqOxMQkiRJkiSp7waWgEiyOMmlSS5O8pNWtn6SM5MsbO+PauVJcliSRUkuSbLDoOKUJEmSJElTb9AtIJ5XVdtX1dPb+IHAWVW1ADirjQO8FFjQXm8EjhhwnJIkSZIkaQpN9y0YOwMntOETgFf0lJ9YnXOB9ZJsMh0BSpIkSZKkh27eANdVwLeTFHBUVf0XsFFV3dCm3whs1IY3A5b0zLu0ld3ACBYuXNifiCWtEjae7gCkGc7PWUmSNBELFiwYc/ogExDPqarrkjwGODPJz3onVlW15MSkjbeRkjSWO6c7AGmG83NWkiRNhYHdglFV17X3m4FTgGcANw3dWtHeb27VrwO26Jl981YmSZIkSZJmoYEkIJKsk+QRQ8PAi4DLgNOAvVq1vYBT2/BpwJ7taRjPAm7vuVVDkiRJkiTNMoO6BWMj4JQkQ+v8fFV9M8kFwBeT7AtcC+zW6p8BvAxYBCwD9h5QnJIkSZIkqQ8GkoCoqquBPxqh/FfATiOUF3DAAEKTJEmSJEkDMN2P4ZQkSZIkSasAExCSJEmSJKnvTEBIkiRJkqS+MwEhSZIkSZL6blBPwZi1nvyp86c7BGnGunz/Z0x3CJIkSZJmCVtASJIkSZKkvjMBIUmSJEmS+s4EhCRJkiRJ6jsTEJIkSZIkqe9MQEiSJEmSpL4zASFJkiRJkvrOBIQkSZIkSeo7ExCSJEmSJKnvTEBIkiRJkqS+MwEhSZIkSZL6zgSEJEmSJEnqOxMQkiRJkiSp70xASJIkSZKkvhtoAiLJ6kkuSnJ6G986yXlJFiX5QpKHtfI12/iiNn2rQcYpSZIkSZKm1qBbQPwjcGXP+IeBj1XVtsCvgX1b+b7Ar1v5x1o9SZIkSZI0Sw0sAZFkc+AvgKPbeIDnA19uVU4AXtGGd27jtOk7tfqSJEmSJGkWmjfAdX0ceCfwiDa+AXBbVS1v40uBzdrwZsASgKpanuT2Vv+XIy144cKF/YpZ0hjmyrG38XQHIM1wc+VYlyRJ/bVgwYIxpw8kAZHk5cDNVfXTJDtO9fLH28iH5Mzz+7dsaZbr67E3QHdOdwDSDDdXjnVJkjS9BtUC4k+Bv0ryMmAt4JHAJ4D1ksxrrSA2B65r9a8DtgCWJpkHzAd+NaBYJUmSJEnSFBtIHxBV9a6q2ryqtgJ2B75bVXsA3wN2adX2Ak5tw6e1cdr071ZVDSJWSZIkSZI09Qb9FIzh/hl4W5JFdH08HNPKjwE2aOVvAw6cpvgkSZIkSdIUGGQnlABU1dnA2W34auAZI9S5F9h1oIFJkiRJkqS+me4WEJIkSZIkaRVgAkKSJEmSJPWdCQhJkiRJktR3JiAkSZIkSVLfmYCQJEmSJEl9ZwJCkiRJkiT1nQkISZIkSZLUdxNOQCR5xyjlb5u6cCRJkiRJ0lw0mRYQ7xml/KCpCESSJEmSJM1d88arkOT5bXD1JM8D0jP5ccCd/QhMkiRJkiTNHeMmIIBj2vtawLE95QXcCLx5qoOSJEmSJElzy7gJiKraGiDJiVW1Z/9DkiRJmnrXnOFljDSarV924nSHIGkVMJEWEAD0Jh+SrDZs2gNTGZQkSZIkSZpbJvMUjB2S/DjJ3cBv2mt5e5ckSZIkSRrVhFtAACcAXwf2AZb1JxxJkiRJkjQXTSYBsSXw7qqqfgUjSZIkSZLmpgnfggGcAryoX4FIkiRJkqS5azItINYCTknyQ7rHb/6OT8eQJEmSJEljmUwC4or2mrQkawE/ANZs6/xyVb03ydbAycAGwE+B11XV/UnWBE4Engb8CnhVVS1emXVLkiRJkqTpN5nHcP7bQ1jPfcDzq+quJGsAP0zy38DbgI9V1clJjgT2BY5o77+uqm2T7A58GHjVQ1i/JEmSJEmaRhNOQCR5/mjTquq7Y83bOq68q42u0V4FPB94TSs/ATiYLgGxcxsG+DJweJLYAaYkSZIkSbPTZG7BOGbY+IbAw4ClwOPGmznJ6nS3WWwLfBL4BXBbVS1vVZYCm7XhzYAlAFW1PMntdLdp/HKkZS9cuHASmyFpqsyVY2/j6Q5AmuHmyrE+mYseaVUzV45zSdNrwYIFY06fzC0YW/eOt4TCQcCdE5z/t8D2Sdaje6LGEye67vGMt5EPyZnn92/Z0izX12NvgCZ0EpNWYXPlWL/G71fSqObKcS5pZpvMYzgfpCUUDgHeOcn5bgO+BzwbWC/JUBJkc+C6NnwdsAVAmz6frjNKSZIkSZI0C610AqJ5IfDAeJWSbNhaPpDk4W2+K+kSEbu0ansBp7bh09o4bfp37f9BkiRJkqTZazKdUC6h6zhyyNrAWsD+E5h9E+CEdtvGasAXq+r0JFcAJyf5AHARK/qZOAb4TJJFwK3A7hONU5IkSZIkzTyT6Y/ptcPG7wZ+XlV3jDdjVV0CPHWE8quBZ4xQfi+w6yRikyRJkiRJM9hkOqH8PkCS1YCNgJuqatzbLyRJkiRJkibcB0SSRyQ5EbiHrpPIe5KckGR+36KTJEmSJElzwmQ6ofxPYB3gKcDD2/vawGF9iEuSJEmSJM0hk+kD4iXA46pqWRv/eZK9gV9MfViSJEmSJGkumUwLiHuBDYeVPRq4b+rCkSRJkiRJc9FkWkAcDZyZ5KPAtcCWwFuBT/cjMEmSJEmSNHdMJgFxCF3nk3sAmwLXAx+pqmP6EZgkSZIkSZo7JnMLxieAq6rqBVW1XVW9ALgyycf7FJskSZIkSZojJpOAeDXwk2FlPwVeM3XhSJIkSZKkuWgyCYgCVh9WtvoklyFJkiRJklZBk0kenAO8P8lqAO394FYuSZIkSZI0qsl0QvmPwOnADUmuBR4L3AD8ZT8CkyRJkiRJc8eEExBVtTTJDsAzgC2AJcD5VfVAv4KTJEmSJElzw2RaQNCSDee2lyRJkiRJ0oTYgaQkSZIkSeo7ExCSJEmSJKnvTEBIkiRJkqS+MwEhSZIkSZL6biAJiCRbJPlekiuSXJ7kH1v5+knOTLKwvT+qlSfJYUkWJbmkPX1DkiRJkiTNUoNqAbEceHtVbQc8CzggyXbAgcBZVbUAOKuNA7wUWNBebwSOGFCckiRJkiSpDwaSgKiqG6rqwjZ8J3AlsBmwM3BCq3YC8Io2vDNwYnXOBdZLsskgYpUkSZIkSVNv3qBXmGQr4KnAecBGVXVDm3QjsFEb3gxY0jPb0lZ2AyNYuHBhP0KVNI65cuxtPN0BSDPcXDnWB37RI80ic+U4lzS9FixYMOb0gX4WJ1kX+Arwlqq6I8nvplVVJamVWe54G/mQnHl+/5YtzXJ9PfYG6M7pDkCa4ebKsX6N36+kUc2V41zSzDawp2AkWYMu+fC5qvpqK75p6NaK9n5zK78O2KJn9s1bmSRJkiRJmoUG9RSMAMcAV1bVR3smnQbs1Yb3Ak7tKd+zPQ3jWcDtPbdqSJIkSZKkWWZQt2D8KfA64NIkF7eyfwE+BHwxyb7AtcBubdoZwMuARcAyYO8BxSlJkiRJkvpgIAmIqvohkFEm7zRC/QIO6GtQkiRJkiRpYAbWB4QkSZIkSVp1mYCQJEmSJEl9ZwJCkiRJkiT1nQkISZIkSZLUdyYgJEmSJElS35mAkCRJkiRJfWcCQpIkSZIk9Z0JCEmSJEmS1HcmICRJkiRJUt+ZgJAkSZIkSX1nAkKSJEmSJPWdCQhJkiRJktR3JiAkSZIkSVLfmYCQJEmSJEl9ZwJCkiRJkiT1nQkISZIkSZLUdyYgJEmSJElS35mAkCRJkiRJfTeQBESSY5PcnOSynrL1k5yZZGF7f1QrT5LDkixKckmSHQYRoyRJkiRJ6p9BtYA4HnjJsLIDgbOqagFwVhsHeCmwoL3eCBwxoBglSZIkSVKfzBvESqrqB0m2Gla8M7BjGz4BOBv451Z+YlUVcG6S9ZJsUlU3DCJWSZIkSbPTe478u+kOQZrR3rffUdO6/oEkIEaxUU9S4UZgoza8GbCkp97SVjZqAmLhwoV9CVDS2ObKsbfxdAcgzXBz5VifzoseaaabK8e5pLH1+1hfsGDBmNNnxGdxVVWSWtn5x9vIh+TM8/u3bGmW6+uxN0B3TncA0gw3V471a/x+JY1qrhznnDXdAUgz23Qf69P5FIybkmwC0N5vbuXXAVv01Nu8lUmSJEmSpFlqOhMQpwF7teG9gFN7yvdsT8N4FnC7/T9IkiRJkjS7DeQWjCQn0XU4+egkS4H3Ah8CvphkX+BaYLdW/QzgZcAiYBmw9yBilCRJkiRJ/TOop2C8epRJO41Qt4AD+huRJEmSJEkapOm8BUOSJEmSJK0iTEBIkiRJkqS+MwEhSZIkSZL6zgSEJEmSJEnqOxMQkiRJkiSp70xASJIkSZKkvjMBIUmSJEmS+s4EhCRJkiRJ6jsTEJIkSZIkqe9MQEiSJEmSpL4zASFJkiRJkvrOBIQkSZIkSeo7ExCSJEmSJKnvTEBIkiRJkqS+MwEhSZIkSZL6zgSEJEmSJEnqOxMQkiRJkiSp70xASJIkSZKkvpuxCYgkL0lyVZJFSQ6c7ngkSZIkSdLKm5EJiCSrA58EXgpsB7w6yXbTG5UkSZIkSVpZqarpjuH3JHk2cHBVvbiNvwugqj44VOf222+feYFLkiRJkiTmz5+f4WUzsgUEsBmwpGd8aSuTJEmSJEmz0ExNQEiSJEmSpDlk3nQHMIrrgC16xjdvZb8zUnMOSZIkSZI0M83UFhAXAAuSbJ3kYcDuwGnTHJMkSZIkSVpJM7IFRFUtT/Im4FvA6sCxVXX5NIclSZIkSZJW0kxtAUFVnVFVj6+qbarqkOmOR7NHktcnObwNH5zkHdMdkzSaJFsluWwKlrNjktMnOc/xSXZZiXl2nFRwK+ZdnOTRKzPvTDBa/En+KsmB48w76b+PNEhjnYuSvC/JC8aZf0o/b6fymElydpKnT8WypEFIcleflrtjkj/pGZ/0dcAE1zPicpNsmuTLE5h/Srd/qs4BfpZPjRnZAkKSNPMkWX26YwBIMq+qlvd5HaF7VPUD49WtqtPo822Cg9hmaTRV9Z7pjkHSlNgRuAv4n+lYeVVdD0x5wkOzy4xtAaGZJ8k6Sb6R5H+TXJbkVa18pyQXJbk0ybFJ1mzli5N8MMnFSX6SZIck30ryiyT79Sz3n5JckOSSJP82yrpfkuTCtu6zWtn6Sb7W5js3yR+OE/82Sb6Z5KdJzknyxJ7yc1v8H+jNuk4kNmkKzEvyuSRXJvlykrUBkryn/f9dluS/2pdikmyb5DvteLgwyTZtOeu2+X/WljdU/2lJvt/+97+VZJPhAYxzHH84yYXArsDtwP1t2oeSXNGOj/83wjI3SPLtJJcnORpIz7TXJjm/nR+OGkpuJNk3yc/btE9nRWum45McmeQ84CPtfHRsq3dRkp1bvdWT/HvPcft3I+3wJG9r+/WyJG9pZVsluSrJicBlPLgz5CFvbvv80p5zSG+rq1HPJ5P9+6T7xebjSX4C/ONI2yFNsdXbcXd5O3YfDg/+NTPJy9r/8E+THJYH/xq4Xfu/vTrJPwxfeDs+j2/H3aVJ3trKJ3tOG+18NWK5NJtlhGvRjH5NPurncpKtgP2At7bP3j9rk56b5H/acTt0nK+b5Kyez7uhz9it0l2r/N55YgQjLfd3La2SrJ3kiy3eU5Kcl55WCkkOadt3bpKNRtgvf9624+J23D+ilf9zi/l/k3yoZ5Zd010z/Hxo25OsleS4Vv+iJM8bq1xTpKp8+ZrQC/gb4NM94/OBtYAlwONb2YnAW9rwYuDv2/DHgEuARwAbAje18hcB/0X3xWQ14HTgucPWu2Fbx9ZtfP32/p/Ae9vw84GL2/DrgcPb8MHAO9rwWcCCNvxM4Ltt+HTg1W14P+Cuicbmy9dDfQFbAQX8aRs/tud/dv2eep8B/rINnwf8dRteC1ib7leN2+meGrQa8GPgOcAadL90bNjqv4quXx2A4+l+iRjvOH7nCHFvAFxF10oAYL0R6hwGvKcN/0XbzkcDTwK+DqzRpn0K2BPYtK1v/Rb3OT3H8vHtGFy9jR8KvHZo3cDPgXWANwIHtfI1gZ8MnTt64noacGmrvy5wOfDU9rd4AHjWKH+rxcCb2/D+wNFt+PU9cY52PlmZv8/ZwKem+3/U16rxav//y4Ht2/gXe46x4eeKoc/jk4DT2/DB7X95zXac/2roGO9Zx9OAM3vG12vvkzmnjXi+Gq28DZ8NPH2697EvXxN9Mc61KCNfk0/kc/lg2jVGGz8e+FJb9nbAolY+D3hkG340sKjFMOp5Yth6RlvuVsBlbfgdwFFt+A/acp/exosV1zwfoX2uD1vH11lx7bRui/ml7Ty0disf+s5wNvAfbfhlwHfa8NtZ8Zn7ROD/2rlktPIdaec8Xyv/sgWEJuNS4IXpfg39s6q6HXgCcE1V/bzVOYHuxDjktJ55z6uqO6vqFuC+JOvRnVhfBFwEXEh3kC8Ytt5nAT+oqmsAqurWVv4cui9lVNV3gQ2SPHKkwJOsC/wJ8KUkFwNHAUO/Aj+b7iQJ8Pme2SYSmzQVllTVj9rwZ+n+twGe134RuJQuyfbkluHfrKpOAaiqe6tqWat/flUtre62gYvpPuifQPfBfmb73z+I7oK+13jH8RdGiPl24F7gmCSvBJaNUOe5bXuoqm8Av27lO9F9EbmgxbQT8DjgGcD3q+rWqvoNK47LIV+qqt+24RcBB7b5z6a7MHhsK9+zlZ9Hd0E2/Lh9DnBKVd1dVXcBXwWGfgm6tqrOHWFbhny1vf+Ubv8ON9r5BFbu7zPSvpf65ZqqurgNj/Q//kTg6qHPY7oERK9vVNV9VfVL4GZg+K+WVwOPS/KfSV4C3LGS57SRzlfjncek2Wi0a9GRrskn8rk8kq9V1QNVdQUrjtkAhya5BPgOsFnPtPHOE2Mtt9dzgJMBquoyuh8qh9xPl2wZax0/Aj7aWlutV91tii8Ajhs6h/R8Z4CRP7+fw4rrlJ8B1wKPH6NcU8A+IDRhVfXzJDvQZQ4/kO5WiFPHme2+9v5Az/DQ+Dy6E9wHq+qoqY53mNWA26pq+0nMM6jYpBo+nmQtupYBT6+qJUkOpvuSPZbeY+y3rDjGLq+qZz+E+O4eXlDd04qeQZc82AV4E12SZCICnFBV73pQYfKKScQR4G+q6qphywhdK4VvTTCWsdYxkqF9PLR/J2Nl/j7jxSNNpeH/o6M1rZ7o/A86Rqrq10n+CHgxXQuh3Rj79qIxlyetAka9Fh1+TV5V71vJz+Xe42zoVsk96FogP62qfpNkMSuuQSZ6nhhpuRP1m6oaujYa8divqg8l+QbdPvhRkhePs8yH8vmtKWQLCE1Ykk2BZVX1WeDfgR3omnptlWTbVu11wPcnsdhvAfu0Fgok2SzJY4bVOZfuPrKtW531W/k5dCdI0vXK/8uqumOklbTya5Ls2uqnXQQNLf9v2vDuk4xNmgqPTTL0BfQ1wA9Z8UH/y/Y/uAtAVd0JLB36sp5kzbQ+I0ZxFbDh0PKTrJHkySPUmdRx3GKaX1VnAG8F/miEaj9o20OSlwKPauVnAbsMHU/p+nPZErgA+PMkj0oyjxXH5Ui+Rdcfw9A94U/tKf/7JGu08scnWWfYvOcAr2j3n64D/HUrmwqjnU9GM5G/jzRTXEXXgmGrNv6qycyc7ikyq1XVV+ha++ywkue0kc5XD/V6RJqJRrwWHemafIKfy3fS3Q49nvnAzS358Dxgy6nYmGF+RJeEJMl2wFMmM3OSbarq0qr6MN31wxOBM4G9s6IvrfXHWgYP/i7xeLqWlFeNUa4pYPZHk/EU4N+TPAD8hq5/h3uT7E13a8M8uhPAkRNdYFV9O8mTgB+37xF3Aa+la7o5VOeWJG8EvppktTbthXT3sR3bmoctA/YaZ3V7AEckOYjuvuuTgf+lu3f0s0neDXyTrgnbhGKTpshVwAFJjgWuAI6oqmVJPk3XGeKNdMfWkNcBRyV5H92xuOtoC66q+9N1/nRYkvl05/2P0/V7MFRnZY7jRwCntpYaAd42Qp1/A05KcjndPZn/19Z3RTsOv92O6d8AB1TVuUkOBc4HbgV+RjseR/D+th2XtGVcA7wcOJquaeWFLTlxC/CglhVVdWFfRMkyAAAGK0lEQVSS49t6oOvL4aKeL1UPxYjnk9FM5O8jzRRVdU+S/YFvJrmbB5+XJmIz4Lh2zAIMtYKazDltxPNVVd33UK5HpJlojGvRbRl2Tc7EPpe/Dnw5XaeSbx5j1Z8Dvp7uFtCf0H0eT7VPASckuaIt/3LG+cwc5i0tOfJAm/e/23lge+AnSe4HzgD+ZZwYjmjbuRx4fVvGaOWT3UaNYKiTEmmV1bKk91RVJdmdrgO5nac7LmlVlGTdqrqrfYE4ha4TqFOmO66J8nyiua7nGA3wSWBhVX1suuOSNLuke/rVGi2puA1dXxNPqKr7pzk09ZktIKSuM7zD28XUbcA+0xyPtCo7OMkL6G5B+TbwtWmOZ7I8n2iue0OSvYCH0XWMZz9JklbG2sD32i2TAfY3+bBqsAWEJEmSJEnqOzuhlCRJkiRJfWcCQpIkSZIk9Z0JCEmSJEmS1HcmICRJ0oyW5PVJfjgN6z0yyb8Oer2SJM1VJiAkSVpFJTk7yd9OdxwzwUhJjqrar6reP10xSZI015iAkCRpFZNOX68Bkviob0mS9CAmICRJmkWSLE7yriRXJPl1kuOSrJXkUUlOT3JLKz89yeY9852d5JAkPwKWAZ8B/gw4PMldSQ5v9SrJfkkWJrktySeTpGc5+yS5sq3jW0m27JlWSQ5IshBYOM52PDHJmUluTXJVkt16pm2Q5LQkdyQ5H9imZ9pWbT3zesoe1JIjyRtajHe2/bRDKz8wyS96yv+6lT8JOBJ4dtsXt7Xy45N8YNhyF7WYT0uy6bBtH3W/SZIkExCSJM1GewAvpvti/njgILrP9OOALYHHAvcAhw+b73XAG4FHAK8HzgHeVFXrVtWbeuq9HPhj4A+B3dq6SLIz8C/AK4EN2/wnDVvHK4BnAtuNFnySdYAzgc8DjwF2Bz6VZGieTwL3ApsA+7TXhCTZFTgY2BN4JPBXwK/a5F/QJV3mA/8GfDbJJlV1JbAf8OO2L9YbYbnPBz5Itz82Aa4FTh5WbcT9JkmSOiYgJEmafQ6vqiVVdStwCPDqqvpVVX2lqpZV1Z2t/M+HzXd8VV1eVcur6jdjLP9DVXVbVf0f8D1g+1a+H/DBqrqyqpYDhwLb97aCaNNvrap7xlj+y4HFVXVci+Ui4CvArklWB/4GeE9V3V1VlwEnTGy3APC3wEeq6oLqLKqqawGq6ktVdX1VPVBVX6BrpfGMCS53D+DYqrqwqu4D3kXXYmKrnjqj7TdJkoQJCEmSZqMlPcPXApsmWTvJUUmuTXIH8ANgvfaFfqT5xnJjz/AyYN02vCXwiXaLwW3ArUCAzSa5ji2BZw4tpy1rD2BjupYV80bYxonagq6lw+9JsmeSi3vW+QfAoye43E1746iqu+haVvRu+2j7TZIk0X3AS5Kk2WWLnuHHAtcDbweeADyzqm5Msj1wEV2CYEgNW87w8fEsAQ6pqs+NUWciy1wCfL+qXjh8QkuYLKfbxp+14sf2VLm7va8N3NGGNx627G0YprXS+DSwE92tFr9NcjEr9s94cV9PlzgZWt46wAbAdePMJ0mSGltASJI0+xyQZPMk6wPvBr5A16/DPcBtrfy9E1jOTcDjJrHeI4F3JXkyQJL5rc+FyTodeHyS1yVZo73+OMmTquq3wFeBg1urju2AvYZmrKpb6L70vzbJ6kn24cEJh6OBdyR5Wnvax7Yt+bAOXZLhlhb73nQtIIbcBGye5GGjxHwSsHeS7ZOsSXf7yXlVtXgltl+SpFWSCQhJkmafzwPfBq6mu93gA8DHgYcDvwTOBb45geV8AtilPdHisPEqV9UpwIeBk9ttHpcBL51s8K2PihfRdT55Pd2tCx8G1mxV3kR3+8KNwPF0nWv2egPwT3S3QDwZ+J+eZX+Jrv+LzwN3Al8D1q+qK4D/AH5Ml2x4CvCjnmV+F7gcuDHJL0eI+TvAv9L1VXEDXdJj98luuyRJq7JUTbb1pSRJmi5JFgN/274QS5IkzRq2gJAkSZIkSX1nJ5SSJGnKJfkz4L9HmlZVPh1CkqRVkLdgSJIkSZKkvvMWDEmSJEmS1HcmICRJkiRJUt+ZgJAkSZIkSX1nAkKSJEmSJPWdCQhJkiRJktR3/x9dvSqMpArDbgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693d3eef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693c9bfd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693cd7438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693c1f358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693c37390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABCYAAADgCAYAAADfVWeFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAF+BJREFUeJzt3XuwJ2V5J/Dv44yARjMgJsTAKBhPNouXIKtoNHFVvABxhSqvaCIgibHE27quwUup8bLrJRtFS9w1chkIpRijCzEmOIuKmgoXQZerMCMGmXEQFRhZAeXy7B+nR4+TmWHOzJnT58f5fKpOne6n3+5+miromS/9dld3BwAAAGAM9xq7AQAAAGDxEkwAAAAAoxFMAAAAAKMRTAAAAACjEUwAAAAAoxFMAAAAAKMRTAAAAACjEUwAAPOiqk6uqq6qz2xi26HDtjvG6A0AGI9gAgCYT99N8qyq2mOj+p8luWaEfgCAkQkmAID5tCrJuUmO3FCoqgcneXqSk2bUfrWqTqqq66rqp1V1bVX91YztT6+qL1fVDVW1vqrOqaoD5vE6AIA5IpgAAObbx5L8SVXVsP4nSc7OLz8x8a4k+yc5NMlUkhckuWLG9vslOT7J7yV5QqYDj3+qqt13bOsAwFyr7h67BwBgEaiqk5PsleRZSdYmeW6Sr2Q6kHh1kl9N8vHuXlpVZyS5sbuP3Mpj3yvJj5K8srtPm/vuAYAdxRMTAMC86u7bkpya5E+T/GGSpUn+fqNhxyd5blVdWlXHVdXBQ/iQJKmqfarq1KpaXVU/TvLjJMuSPGR+rgIAmCtLx24AAFiUPpbkoiTLk5zU3bf/YmZH0t1nDe+eeGaSJyf5mySXVNWB3X1nks8l+WGSY5Jcm+RnSb6WZKf5vAgAYPsJJgCAedfdl1fVBUmemBkvwtxozA1JPpHkE1V1UpJ/SbJvVX0vyb5JDunus5KkqvZK8uvz0TsAMLcEEwDAWJ6ZZJchgPglVfXuJBcmuSzJXUlenOT/Zfpzozcn+UGSP62qbyfZPcn7ktw6T30DAHPIOyYAgFF09y2bCiUGtyV5R6bDia8neVSSg7t7fXffleR5SX4rycVJTk7ywSTrdnjTAMCc81UOAAAAYDSemAAAAABGI5gAAAAARiOYAAAAAEZzj/sqx/r16700AwAAABagZcuW1cY1T0wAAAAAoxFMAAAAAKMRTAAAAACjEUwAAAAAoxFMAAAAAKMRTAAAAACjEUwAAAAAoxFMAAAAAKNZOnYDbN7Djz9/7BYAmFCXveKAsVsAANgqnpgAAAAARiOYAAAAAEYjmAAAAABGI5gAAAAARiOYAAAAAEYzL8FEVZ1YVddX1aUzau+vqm9V1cVV9dmq2nXGtjdW1eqqurKqnjmjftBQW11Vx85H7wAAAMCOM19PTJyc5KCNaiuTPKK7H5XkqiRvTJKq2jfJC5M8fNjn+KpaUlVLknwkycFJ9k1y+DAWAAAAmFDzEkx091eS3LBR7Qvdfcewem6SvYblQ5N8srt/2t3fSbI6yQHDz+ruvrq7f5bkk8NYAAAAYEItHbuBwUuTnD4s75npoGKDNUMtSa7dqP64LR101apVc9UfAEwU90AAYKGYmpra4vbRg4mqenOSO5KcNtfHvruLX/BWnj92BwBMqIm/BwIAi8aowURVHZnkWUkO7O4eymuTLJ8xbK+hli3UAQAAgAk02udCq+qgJG9I8uzuvmXGpjOTvLCqdq6qfZJMJTk/yQVJpqpqn6raKdMvyDxzvvsGAAAA5s68PDFRVZ9I8uQkD6yqNUnelumvcOycZGVVJcm53f3y7r6sqj6V5PJMT/E4prvvHI7zyiRnJVmS5MTuvmw++gcAAAB2jPrFDIp7hvXr199jLujhx3vHBADb5rJXHDB2CwAA/8ayZctq49poUzkAAAAABBMAAADAaAQTAAAAwGgEEwAAAMBoBBMAAADAaAQTAAAAwGgEEwAAAMBoBBMAAADAaAQTAAAAwGgEEwAAAMBoBBMAAADAaAQTAAAAwGgEEwAAAMBoBBMAAADAaAQTAAAAwGjmJZioqhOr6vqqunRG7QFVtbKqVg2/dxvqVVUfqqrVVXVxVe0/Y58jhvGrquqI+egdAAAA2HHm64mJk5MctFHt2CRnd/dUkrOH9SQ5OMnU8POyJB9NpoOMJG9L8rgkByR524YwAwAAAJhM8xJMdPdXktywUfnQJCuG5RVJDptRP6WnnZtk16p6UJJnJlnZ3Td0941JVubfhh0AAADABFk64rn36O51w/J1SfYYlvdMcu2McWuG2ubqm7Vq1aq56RQAJox7IACwUExNTW1x+5jBxM91d1dVz/Vx7+7iF7yV54/dAQATauLvgQDAojHmVzm+P0zRyPD7+qG+NsnyGeP2GmqbqwMAAAATasxg4swkG76scUSSM2bUXzJ8nePxSdYPUz7OSvKMqtpteOnlM4YaAAAAMKHmZSpHVX0iyZOTPLCq1mT66xrvSfKpqjo6yTVJnj8M/3ySQ5KsTnJLkqOSpLtvqKp3JrlgGPeO7t74hZoAAADABJmXYKK7D9/MpgM3MbaTHLOZ45yY5MQ5bA0AAAAY0ZhTOQAAAIBFTjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMZvRgoqr+c1VdVlWXVtUnqmqXqtqnqs6rqtVVdXpV7TSM3XlYXz1s33vc7gEAAIDtMWowUVV7Jnl1ksd09yOSLEnywiTvTfKB7n5YkhuTHD3scnSSG4f6B4ZxAAAAwIQa/YmJJEuT3Keqlia5b5J1SZ6a5NPD9hVJDhuWDx3WM2w/sKpqHnsFAAAA5tDSMU/e3Wur6i+TfDfJrUm+kOTCJDd19x3DsDVJ9hyW90xy7bDvHVW1PsnuSX64qeOvWrVqB3YPAAuXeyAAsFBMTU1tcfuowURV7ZbppyD2SXJTkr9NctBcHf/uLn7BW3n+2B0AMKEm/h4IACwaWz2Vo6pev5n667bj/E9L8p3u/kF3357kM0memGTXYWpHkuyVZO2wvDbJ8uG8S5MsS/Kj7Tg/AAAAMKLZvGPirZupv2U7zv/dJI+vqvsO74o4MMnlSb6U5LnDmCOSnDEsnzmsZ9j+xe7u7Tg/AAAAMKK7ncpRVU8dFpdU1VOSzHzZ5EOT3LytJ+/u86rq00kuSnJHkm8k+ViSf0jyyap611A7YdjlhCSnVtXqJDdk+gseAAAAwITamndMbAgFdkly4ox6J7kuyau2p4HufluSt21UvjrJAZsYe1uS523P+QAAAICF426Die7eJ0mq6pTufsmObwkAAABYLLb6qxwzQ4mqutdG2+6ay6YAAACAxWE2X+XYv6r+pap+kuT24eeO4TcAAADArG31ExNJViT5+yQvTXLLjmkHAAAAWExmE0w8JMmbfZ4TAAAAmCtbPZUjyWeTPGNHNQIAAAAsPrN5YmKXJJ+tqq9l+jOhP+drHQAAAMC2mE0wcfnwAwAAADAnZvO50L/YkY0AAAAAi89WBxNV9dTNbevuL85NOwAAAMBiMpupHCdstP5rSXZKsibJQ+esIwAAAGDRmM1Ujn1mrlfVkiRvSXLzXDcFAAAALA6z+VzoL+nuO5O8O8kb5q4dAAAAYDHZ5mBi8PQkd81FIwAAAMDiM5uXX16bpGeU7ptklySvmOumAAAAgMVhNi+//KON1n+S5Kru/vH2NFBVuyb5eJJHZDr4eGmSK5OcnmTvJP+a5PndfWNVVZLjkhyS5JYkR3b3RdtzfgAAAGA8Wz2Vo7vP6e5zknw1yVVJLtreUGJwXJJ/6u7fSfK7Sa5IcmySs7t7KsnZw3qSHJxkavh5WZKPzsH5AQAAgJFsdTBRVfevqlOS3JpkbZJbq2pFVS3b1pMP+z4pw6dIu/tn3X1TkkOTrBiGrUhy2LB8aJJTetq5SXatqgdt6/kBAACAcc1mKseHk/xKkkcmuSbJQzL9VY4PJTliG8+/T5IfJDmpqn43yYVJXpNkj+5eN4y5Lskew/KeSa6dsf+aobYum7Bq1aptbAsAJpt7IACwUExNTW1x+2yCiYOSPLS7bxnWr6qqo5J8ext723D+/ZO8qrvPq6rj8otpG0mS7u6q6k3ufTfu7uIXvJXnj90BABNq4u+BAMCiMZvPhd6W5Nc2qj0wyU+34/xrkqzp7vOG9U9nOqj4/oYpGsPv64fta5Msn7H/XkMNAAAAmECzCSY+nmRlVb28qg6uqpcnOSvJX2/rybv7uiTXVtW/G0oHJrk8yZn5xfSQI5KcMSyfmeQlNe3xSdbPmPIBAAAATJjZTOV4d6afTnhxkt9M8r0k7+vuE7azh1clOa2qdkpydZKjMh2YfKqqjs70+yyeP4z9fKY/Fbo6058LPWo7zw0AAACMaDbBxHFJPtndT9tQqKonVNUHu/u129pAd38zyWM2senATYztJMds67kAAACAhWU2UzkOT/L1jWoXJnnR3LUDAAAALCazCSY6yZKNaktmeQwAAACAn5tNqPDVJO+sqnslyfD77UMdAAAAYNZm846J1yT5XJJ1VXVNkgcnWZfkP+2IxgAAAIB7vq0OJrp7TVXtn+SAJMuTXJvk/O6+a0c1BwAAANyzzeaJiQwhxLnDDwAAAMB28eJKAAAAYDSCCQAAAGA0ggkAAABgNIIJAAAAYDSCCQAAAGA0ggkAAABgNIIJAAAAYDSCCQAAAGA0ggkAAABgNAsimKiqJVX1jar63LC+T1WdV1Wrq+r0qtppqO88rK8etu89Zt8AAADA9lkQwUSS1yS5Ysb6e5N8oLsfluTGJEcP9aOT3DjUPzCMAwAAACbU6MFEVe2V5A+TfHxYryRPTfLpYciKJIcNy4cO6xm2HziMBwAAACbQ0rEbSPLBJG9Icv9hffckN3X3HcP6miR7Dst7Jrk2Sbr7jqpaP4z/4aYOvGrVqh3VMwAsaO6BAMBCMTU1tcXtowYTVfWsJNd394VV9eS5Pv7dXfyCt/L8sTsAYEJN/D0QAFg0xn5i4olJnl1VhyTZJcmvJjkuya5VtXR4amKvJGuH8WuTLE+ypqqWJlmW5Efz3zYAAAAwF0Z9x0R3v7G79+ruvZO8MMkXu/vFSb6U5LnDsCOSnDEsnzmsZ9j+xe7ueWwZAAAAmEOjv/xyM/48yeuqanWm3yFxwlA/IcnuQ/11SY4dqT8AAABgDow9lePnuvvLSb48LF+d5IBNjLktyfPmtTEAAABgh1moT0wAAAAAi4BgAgAAABiNYAIAAAAYjWACAAAAGI1gAgAAABiNYAIAAAAYjWACAAAAGI1gAgAAABiNYAIAAAAYjWACAAAAGI1gAgAAABiNYAIAAAAYjWACAAAAGI1gAgAAABiNYAIAAAAYjWACAAAAGM2owURVLa+qL1XV5VV1WVW9Zqg/oKpWVtWq4fduQ72q6kNVtbqqLq6q/cfsHwAAANg+Yz8xcUeS/9Ld+yZ5fJJjqmrfJMcmObu7p5KcPawnycFJpoaflyX56Py3DAAAAMyVUYOJ7l7X3RcNyzcnuSLJnkkOTbJiGLYiyWHD8qFJTulp5ybZtaoeNM9tAwAAAHNk6dgNbFBVeyd5dJLzkuzR3euGTdcl2WNY3jPJtTN2WzPU1mUTVq1atSNaBYAFzz1wYfmN4986dgsATKjrXvGOsVvYblNTU1vcviCCiaq6X5K/S/La7v5xVf18W3d3VfW2HPfuLn7BW3n+2B0AMKEm/h54D3Pz2A0AMLEWwz197HdMpKrunelQ4rTu/sxQ/v6GKRrD7+uH+toky2fsvtdQAwAAACbQ2F/lqCQnJLmiu/9qxqYzkxwxLB+R5IwZ9ZcMX+d4fJL1M6Z8AAAAABNm7KkcT0zyx0kuqapvDrU3JXlPkk9V1dFJrkny/GHb55MckmR1kluSHDW/7QIAAABzadRgoru/lqQ2s/nATYzvJMfs0KYAAACAeTP6OyYAAACAxUswAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMRjABAAAAjEYwAQAAAIxGMAEAAACMZiKDiao6qKqurKrVVXXs2P0AAAAA22bigomqWpLkI0kOTrJvksOrat9xuwIAAAC2RXX32D3MSlX9XpK3d/czh/U3Jkl3//ckWb9+/WRdEAAAACwSy5Ytq41rE/fERJI9k1w7Y33NUAMAAAAmzCQGEwAAAMA9xNKxG9gGa5Msn7G+11BLsunHQgAAAICFaRKfmLggyVRV7VNVOyV5YZIzR+4JAAAA2AYT98REd99RVa9MclaSJUlO7O7LRm4LAAAA2AYT91UOgJmqar8kv9ndnx+7FwBgflTVm7r7v43dBzA3JnEqB8BM+yU5ZFMbqmringoDALbKmzZVrGn+jgMTxr+0wOiqau+q+lZVnVxVV1XVaVX1tKr656paVVUHVNWvVNWJVXV+VX2jqg4d3jPzjiQvqKpvVtULqurtVXVqVf1zklOrapeqOqmqLhn2e8rIlwsAE2O4R19RVX9dVZdV1Req6j5VtV9VnVtVF1fVZ6tqt2H8l6vqvcP9+qqq+oPNHPfLVfWBqvr6cPzHVtVnhvv+u2aM+6PhWN+sqv9VVUuq6j1J7jPUTht6vLKqTklyaZLlVXX4cO+/tKreOy//sIBtJpgAFoqHJfkfSX5n+HlRkt9P8vpM/1+RNyf5YncfkOQpSd6f5N5J3prk9O7er7tPH461b5KndffhSY5J0t39yCSHJ1lRVbvM32UBwMSbSvKR7n54kpuSPCfJKUn+vLsfleSSJG+bMX7pcL9+7Ub1jf2sux+T5H8mOSPT9+xHJDmyqnavqn+f5AVJntjd+yW5M8mLu/vYJLcO9/4Xz+jx+KHH25O8N8lTM/1k5WOr6rDt/8cA7CgecwYWiu909yVJUlWXJTm7u7uqLkmyd6Y/Dfzsqnr9MH6XJA/ezLHO7O5bh+XfT/LhJOnub1XVNUl+O8nFO+YyAOAe5zvd/c1h+cIkv5Vk1+4+Z6itSPK3M8Z/ZsbYvbdw3A1f1rskyWXdvS5JqurqJMszfQ//D0kuqKokuU+S6zdzrGu6+9xh+bFJvtzdPxiOd1qSJyX531u+TGAsgglgofjpjOW7Zqzflen/Vt2Z5DndfeXMnarqcZs41k92SIcAsDjNvEffmWTXrRx/Z4a/b1TVSUkeneR73X3IRuNm3vc3rC9NUklWdPcbt6JH936YYKZyAJPirCSvquF/mVTVo4f6zUnuv4X9vprkxcM+v53ppyyu3MJ4AGDL1ie5ccb7I/44yTlbGJ/uPmqYerHJF1ZvxtlJnltVv54kVfWAqnrIsO32qrr3ZvY7P8l/rKoHVtWSTE/l3GJ/wLgEE8CkeGem3ylx8TDV451D/UtJ9t3w8stN7Hd8knsNU0JOT3Jkd/90E+MAgK13RJL3V9XFmX6Pwzvm+gTdfXmStyT5wnCelUkeNGz+WKb/THDaJvZbl+TYTP8Z4f8mubC7z5jr/oC5U909dg8AAADAIuWJCQAAAGA0ggkAAABgNIIJAAAAYDSCCQAAAGA0ggkAAABgNIIJAAAAYDSCCQAAAGA0ggkAAABgNIIJAGDeVNW/VtV/raqLq+onVXVCVe1RVf9YVTdX1f+pqt2qapeq+puq+lFV3VRVF1TVHsMxjqqqK4bxV1fVn419XQDAtls6dgMAwKLznCRPz/SfQ76R5NFJjk5yRZLPJ3l1kuuSLEuyPMlPk+yX5NZh/+uTPCvJ1UmelOQfq+qC7r5oHq8BAJgjnpgAAObbh7v7+929NslXk5zX3d/o7tuSfDbTQcXtSXZP8rDuvrO7L+zuHydJd/9Dd3+7p52T5AtJ/mCkawEAtpNgAgCYb9+fsXzrJtbvl+TUJGcl+WRVfa+q3ldV906Sqjq4qs6tqhuq6qYkhyR54Dz1DgDMMcEEALDgdPft3f0X3b1vkidkeurGS6pq5yR/l+Qvk+zR3btmevpHjdctALA9BBMAwIJTVU+pqkdW1ZIkP8701I67kuyUZOckP0hyR1UdnOQZ43UKAGwvwQQAsBD9RpJPZzqUuCLJOUlO7e6bM/1yzE8luTHJi5KcOVaTAMD2q+4euwcAAABgkfLEBAAAADAawQQAAAAwGsEEAAAAMBrBBAAAADAawQQAAAAwGsEEAAAAMBrBBAAAADAawQQAAAAwmv8PE+BuMrdOUqkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56939261d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693862630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f569386d198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687fb6518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687f47128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "## Numeric Features"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693be4da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAADgCAYAAAA31tA3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAFQlJREFUeJzt3X+QZWV95/H3h8GfaIYgMswCOhJbXCBxVIqV0iUoJIJrolIbArUbQEnEKt2QhFQCxFqILu4vkTWVRAmBDNkYAgkQCcsaCZsKbiqYgLA6/MoFQX449IBgY5BQMHz3j3N6uLQDPXPv7dN9ut+vqlt9zvOce87z1L23v/d5znOfJ1WFJEkLbafFLoAkaWUw4EiSOmHAkSR1woAjSeqEAUeS1AkDjiSpEwYcSVInDDjSiJLsleTJJN9OsvNil0da6gw40uhOAq4Cvgv81CKXRVryDDjSCJLsRBNwNgAXAR+ek/+qJH+a5PEk00k+meSiJH8157j/kOT2JP+cZJDkN2wtabky4EijOQp4CfC/gf8JHJ5k3VD+HwBvAt4LvAvYG3j/8AmSnAX8KnA68C+BU4CTgTMXtOTSIolzqUk7LskXgTur6tR2/0vADVX18SRTwD8CR1TVtW3+i4BvAndU1RFJXg48DBxdVV8aOu/xwG9V1a4dV0lacDbdpR2UZC/g3wBvHkq+CDinbbXs36ZdP5tZVU8luQF4ZZt0APAy4LIkw9/6VgEvTfLqqnpogaogLQoDjrTjTqIJDDclGU5fRTN44Jl2/4W6D2a7s3+GpjU01yNjllFacgw40g4YGizwKeDiOdln0Awe+MV2/xBgtkttZ+CtPBtcbgH+Gdi3qq5e4GJLS4IBR9oxRwH7AOdV1b3DGUk20AwieAr4C+B3kpwMPAScCvwQbaunqv4pyaeAT7Vdan9F83n8UeDNVfXr3VRH6o6DBqQd0A4W2KOqDtlG3s7AJuA84Nz273uAfwI+TzMS7aVV9VNDz/l54GPAG4EnaFpAG6rqcwtcFalzBhypA0lWAbcDV86ObJNWGrvUpAWQ5FBgD+AmmpFpvwyso/mhqLQiGXCkhbEK+Djwepp7OhuBd1bVNxa1VNIisktNktSJRWnhzMzMGOUkaZlbvXr1c36o5lxqkqROzBtwklyYZHOSjUNplyS5uX3ck+TmNn1dkieG8j6/kIWXJPXH9rRwNgBHDidU1c9W1fqqWg9cBlw+lH3XbF5VfWRyRd0+g8Gg60suKOuzdC2nuoD1WeqWQ33mvYdTVdfNmXZ9qzQTSR1DM/26JEnPa7tGqbUB56qqOnBO+qHAZ6rqoKHjbqH5tfRjwMer6itzzzc8aGA5RG1JUmNqamrr9txBA+OOUjuO505guAl4TVV9J8lbgT9PckBVPbY9hZuEwWAw8XMuJuuzdC2nuoD1WeqWQ31GHqXWzht1NHDJbFpVPVlV32m3bwTuAt4wbiElSf03zrDoI4Dbq+r+2YQkr27njCLJvsAUzSqHkqQVbt4utSQXA4cBuye5Hzizqi4AjuUH1wM5FPhEkqdoFqH6SFW5kJRe0IY7Hp/YuU7cb5eJnUvSZG3PKLXjnif9xG2kXUYzTFqSpOdwpgFJUicMOJKkThhwJEmdMOBIkjphwJEkdcKAI0nqhAFHktQJA44kqRMGHElSJww4kqROGHAkSZ0w4EiSOmHAkSR1woAjSerEvAEnyYVJNifZOJR2VpIHktzcPt4zlHd6kjuT3JHk3QtVcElSv2xPC2cDcOQ20s+tqvXt42qAJPvTLMx2QPuc351dAVSStLLNG3Cq6jpge1ftfB/wJ1X1ZFXdDdwJHDxG+SRJy8S8K36+gI8lOR64ATi1qh4F9gKuHzrm/jbteQ0GgzGK0N05F9Nyr8/05sk1ggc7bZnYubbresv8tek769O9qamp580bNeB8DvgkUO3fc4APjXKiFyrcKAaDwcTPuZhWQn3WPPP4xM4/NbXLxM41n5Xw2vSZ9Vl6RhqlVlXTVbWlqp4BzufZbrMHgH2GDt27TZMkrXAjBZwka4d2PwDMjmC7Ejg2yUuSvA6YAv5+vCJKkpaDebvUklwMHAbsnuR+4EzgsCTrabrU7gFOBqiqW5JcCtwKPA18tKq67VSXJC1J8wacqjpuG8kXvMDxZwNnj1MoSdLy40wDkqROGHAkSZ0w4EiSOmHAkSR1woAjSeqEAUeS1AkDjiSpEwYcSVInDDiSpE4YcCRJnTDgSJI6YcCRJHXCgCNJ6oQBR5LUiXkDTpILk2xOsnEo7b8nuT3J15NckWTXNn1dkieS3Nw+Pr+QhZck9cf2tHA2AEfOSbsGOLCqfgz4R+D0oby7qmp9+/jIZIopSeq7eQNOVV0HPDIn7ctV9XS7ez2w9wKUTZK0jKSq5j8oWQdcVVUHbiPvL4BLquqP2uNuoWn1PAZ8vKq+Mvc5MzMzWy86GAxGLLqWi8sfXDWxcx29pyuaS4tpampq6/bq1asznDfvEtMvJMlvAE8DX2iTNgGvqarvJHkr8OdJDqiqx7ancJMwGAwmfs7FtBLqs+aZxyd2/qmpXSZ2rvmshNemz6zP0jNywElyIvBe4PBqm0lV9STwZLt9Y5K7gDcAN4xfVKm/NtwxuaB64n7dBVVpkkYKOEmOBH4N+PGq+v5Q+quBR6pqS5J9gSngmxMpaY9M8p/L2x24LmmZmDfgJLkYOAzYPcn9wJk0o9JeAlyTBOD6dkTaocAnkjwFPAN8pKoe2eaJJUkryrwBp6qO20byBc9z7GXAZeMWSpK0/Iw1aEBS97wfpL7yDoEkqRO2cKTnMV9LYnrzqokO6ZaWO1s4kqROGHAkSZ0w4EiSOuE9HI1k1JFS3veQVi5bOJKkThhwJEmdsEttBZnkDwYlaUfZwpEkdcKAI0nqhF1qWlbsNpSWru1q4SS5MMnmJBuH0nZLck2SQfv3h9v0JPmtJHcm+XqStyxU4SVJ/bG9XWobgCPnpJ0GXFtVU8C17T7AUTQLr00BHwY+N34xJUl9t10Bp6quA+YupPY+4KJ2+yLg/UPpf1iN64Fdk6ydRGElSf01zj2cNVW1qd1+EFjTbu8F3Dd03P1t2ia2YTAYjFGEbVuIc+6I6c2rJneyPSdXn4mWawzTm6cXuwgT0/e6DHba8tz9Rf7sTJr16d7U1NTz5k1k0EBVVZIa5bkvVLhRDAaDiZ9zR0126pZvT6w+S2FKmenN06zZY838B/bAcqjL1NSzC7Athc/OJFmfpWecYdHTs11l7d/NbfoDwD5Dx+3dpkmSVrBxAs6VwAnt9gnAF4fSj29Hq70NmBnqepMkrVDb1aWW5GLgMGD3JPcDZwL/Bbg0yUnAt4Bj2sOvBt4D3Al8H/jghMssSeqh7Qo4VXXc82Qdvo1jC/joOIWSJC0/Tm0jSeqEAUeS1AkDjiSpEwYcSVInDDiSpE70enmCbU1FP7151Ui/qD9xv13mP0iSNLJeB5xJch0VSVpYdqlJkjphwJEkdcKAI0nqhPdwlrjLHxxtEIQkLTW2cCRJnTDgSJI6YcCRJHVi5Hs4SfYDLhlK2hf4j8CuwC8AD7XpZ1TV1SOXUJK0LIwccKrqDmA9QJJVNMtIX0Gz4Nq5VfXpiZRQkrQsTKpL7XDgrqr61oTOJ0laZtIs0DnmSZILga9V1W8nOQs4EXgMuAE4taoeHT5+ZmZm60UHg8HI1738wVUjP1cSHL3nlsUugpaZqamprdurV6/OcN7YASfJi4FvAwdU1XSSNcDDQAGfBNZW1YeGnzMccMax7ck7p1mzx5pJnH5JsD5L13Koy/CktYPB4Dn/LPrO+iy+uQFnEl1qR9G0bqYBqmq6qrZU1TPA+cDBE7iGJKnnJhFwjgMunt1JsnYo7wPAxglcQ5LUc2NNbZNkF+AngJOHkv9bkvU0XWr3zMmTJK1QYwWcqnoceNWctJ8bq0SSpGXJmQYkSZ0w4EiSOmHAkSR1woAjSeqEAUeS1AkDjiSpEy4xLWkitjXV1KiGp9zR8mELR5LUCQOOJKkTBhxJUie8hyOtYMP3XaY3r2LNM5O7DyPNZQtHktQJA44kqRMGHElSJ8a+h5PkHuB7wBbg6ao6KMluwCXAOpo1cY6pqkfHvZYkqb8m1cJ5Z1Wtr6qD2v3TgGuragq4tt2XJK1gC9Wl9j7gonb7IuD9C3QdSVJPpKrGO0FyN/AozZLS51XV7yX5blXt2uYHeHR2H2BmZmbrRQeDwcjXvvzBVSM/V9LSdfSeWxa7CBrR1NTU1u3Vq1dnOG8Sv8N5R1U9kGQP4Joktw9nVlUled6oNly4HbWt3wxMb55mzR5rRj7nUmN9lq7lVBdYWvWZmhp/LrXBYDDW/5elZjnUZ+wutap6oP27GbgCOBiYTrIWoP27edzrSJL6bayAk2SXJK+c3QZ+EtgIXAmc0B52AvDFca4jSeq/cbvU1gBXNLdp2Bn446r6UpJ/AC5NchLwLeCYMa8jSeq5sQJOVX0TeNM20r8DHD7OuSVJy4uTd0paciaxmNvsZKQu5rZ0OLWNJKkTBhxJUicMOJKkThhwJEmdMOBIkjphwJEkdcKAI0nqhAFHktQJA44kqRMGHElSJww4kqROGHAkSZ0w4EiSOjHybNFJ9gH+kGZNnAJ+r6o+m+Qs4BeAh9pDz6iqq8ctqCSNYhIzT89y5unxjLM8wdPAqVX1tXbVzxuTXNPmnVtVnx6/eJKk5WLkgFNVm4BN7fb3ktwG7DWpgknSUmNraTypqvFPkqwDrgMOBH4FOBF4DLiBphX06PDxMzMzWy86GAxGvu7lD64a+bmStJiO3nPLYhdhQUxNTW3dXr16dYbzxg44SV4B/A1wdlVdnmQN8DDNfZ1PAmur6kPDzxkOOOPY1reN6c3TrNljzSROvyRYn6VrOdUFrE/XdrSFMxgMnvPPvA/mBpyxRqkleRFwGfCFqrocoKqmq2pLVT0DnA8cPM41JEnLw8gBJ0mAC4DbquozQ+lrhw77ALBx9OJJkpaLcUapvR34OeAbSW5u084AjkuynqZL7R7g5LFKKElaFsYZpfZ/gWwjy9/cSJJ+gDMNSJI6YcCRJHXCgCNJ6sQ4gwYkSSPa0VkLpjevYs0z235OX2YtsIUjSeqEAUeS1AkDjiSpE97DkaSe68ss1rZwJEmdMOBIkjphwJEkdcKAI0nqhAFHktSJBQs4SY5MckeSO5OctlDXkST1w4IEnCSrgN8BjgL2p1kjZ/+FuJYkqR9SVZM/aXIIcFZVvbvdPx2gqv4zwMzMzOQvKklaUlavXv2cNdMWqkttL+C+of372zRJ0grloAFJUicWamqbB4B9hvb3btOAH2xmSZKWv4Vq4fwDMJXkdUleDBwLXLlA15Ik9cCCtHCq6ukkHwP+ElgFXFhVtyzEtSRJ/bBg93Cq6uqqekNV/UhVnb0Q10iyT5K/TnJrkluSnNKm75bkmiSD9u8PL8T1JynJS5P8fZL/19blN9v01yX5avt7pkvaFmNvJFmV5KYkV7X7va1PknuSfCPJzUluaNN6916blWTXJH+W5PYktyU5pI/1SbJf+5rMPh5L8kt9rMusJL/c/h/YmOTi9v9Dbz87s/o+aOBp4NSq2h94G/DR9vc+pwHXVtUUcG27v9Q9Cbyrqt4ErAeOTPI24L8C51bV64FHgZMWsYyjOAW4bWi/7/V5Z1Wtr6qD2v0+vtdmfRb4UlW9EXgTzevUu/pU1R3ta7IeeCvwfeAKelgXgCR7Ab8IHFRVB9L0Eh1L/z87UFXL5gF8EfgJ4A5gbZu2Frhjscu2g/V4OfA14F8BDwM7t+mHAH+52OXbgXrsTfNBfxdwFZCe1+ceYPc5ab18rwGrgbtpf4vX9/oMlf8ngb/tc1149mclu9Hc9rgKeHefPzuzj763cLZKsg54M/BVYE1VbWqzHgTWLFKxdkjb/XQzsBm4BrgL+G5VPd0e0rffM/0P4NeAZ9r9V9Hv+hTw5SQ3Jvlwm9bL9xrwOuAh4A/aLs/fT7IL/a3PrGOBi9vtXtalqh4APg3cC2wCZoAb6fdnB+h/lxoASV4BXAb8UlU9NpxXzdeBXsxsUFVbqukW2Bs4GHjjIhdpZEneC2yuqhsXuywT9I6qegvNlE0fTXLocGaf3ms035zfAnyuqt4MPM6cLqee1Yf2nsZPA386N69PdWnvNb2P5kvBvwB2AY5c1EJNSO8DTpIX0QSbL1TV5W3ydJK1bf5amhZDb1TVd4G/pmk275pkdjThc37PtMS9HfjpJPcAf0LTrfZZ+luf2W+eVNVmmnsEB9Pf99r9wP1V9dV2/89oAlBf6wPNF4GvVdV0u9/XuhwB3F1VD1XVU8DlNJ+n3n52ZvU64CQJcAFwW1V9ZijrSuCEdvsEmns7S1qSVyfZtd1+Gc29qNtoAs+/bQ/rRV0Aqur0qtq7qtbRdHP8n6r6d/S0Pkl2SfLK2W2aewUb6eF7DaCqHgTuS7Jfm3Q4cCs9rU/rOJ7tToP+1uVe4G1JXt7+j5t9bXr52Rm2IJN3diXJO4CvAN/g2fsEZ9Dcx7kUeA3wLeCYqnpkUQq5nZL8GHARzYiUnYBLq+oTSfalaSHsBtwE/PuqenLxSrrjkhwG/GpVvbev9WnLfUW7uzPwx1V1dpJX0bP32qwk64HfB14MfBP4IO17j57Vp/0ScC+wb1XNtGl9fm1+E/hZmpG4NwE/T3PPpnefnWG9DjiSpP7odZeaJKk/DDiSpE4YcCRJnTDgSJI6YcCRJHXCgCNJ6oQBR5LUCQOOJKkTBhxpDElOS3JXku+1CwF+oE1fleScJA8nuTvJx5LU7FxYSVYnuSDJpiQPJPlPSVYtbm2khbUgS0xLK8hdwL+mmf7+Z4A/SvJ6mtl+j6JZTO9xfnAG4w00k0m+nmY24Kto1kA5r5NSS4vAqW2kCWrXMzqTZqXTS6rqvDb9CJo1jl5Esy7QvcCuVfVEm38c8OGqeueiFFzqgC0caQxJjgd+BVjXJr0C2J1mHZP7hg4d3n4tTeDZ1EwGDDTd28PHSMuOAUcaUZLXAufTTB//d1W1pW3hhGalxr2HDt9naPs+4Ema5aqfRlohHDQgjW4XmlUkHwJI8kHgwDbvUuCUJHu16xz9+uyT2mWPvwyck+SHkuyU5EeS/Hi3xZe6ZcCRRlRVtwLnAH8HTAM/Cvxtm30+TVD5Os3aJVfTrG2ypc0/nmYdmluBR2lW3FzbVdmlxeCgAakDSY4CPl9Vr13sskiLxRaOtACSvCzJe5LsnGQvmpFrV8z3PGk5s4UjLYAkLwf+Bngj8ATwv4BTquqxRS2YtIgMOJKkTtilJknqhAFHktQJA44kqRMGHElSJww4kqRO/H8WCQy44xMCaAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687fe6908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693cb6dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693ca0470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZwAAADgCAYAAAA31tA3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAF6RJREFUeJzt3XmUXGWZx/HvLwuEYUlYpIUECUrLMiqLrCNqDKMsMgYVGDBCQAScgzMw6mHTc8SFGTLHA8IMKkiEILuskUXghMVhRvYdIlMhBElIugWSsAmS8Mwf922o7q7uru6qejvV9fucU6fvfe9b9z71JNVPv2/dulcRgZmZWaONGu4AzMysNbjgmJlZFi44ZmaWhQuOmZll4YJjZmZZuOCYmVkWLjjWdCTdKen8QfQPSV9tZEy1kDQlxTipDvsaVG7McnLBsdWCpOsl3dfHtnGSXpb049T0JeBbg9j9JsBVZftbKenwIQf73n6i7PGGpKckDSauWo79PUkLK2wabG7MsnHBsdXFecDOkrarsO3LwHjgfICIeDkiXql2xxGxNCLerE+YvXyToqD9LXAWMFPS0Q061oAGmxuznFxwbHVxM/An4KgK244Cbo2IhdB72kjSHpL+R9Kr6fGopL3Ktr87pZZGBaOBC7pGJ6l9PUkXSFoq6S1Jz0s6o4q4V6SC9mxEnAs8BuxV3kHSlpKulrRc0jJJt0r6aF87VOGXkp6R9BdJCyT9m6Q10/bDgR8Bm5eNsE7tmRtJR0laIWlcj/2fKOlPkkZVE18NuTHrxgXHVgsR8Q4wC5guaa2udkntwKcpRkC9SBoDzAHuBXZMj1OBN/o41M7AKuB4ipHJJqn9x+m504B24B+BedXGn4rEnsA2wF/L2tuAu4FO4JPAbsDTwJ2S3tfX7lL/r6T9HQ8cAZyStl8BzAQWlb2Gn1TYz5XAGuk1lTsMuDgi3qkyvppyY/auiPDDj9XiAUwEVgKHlbXNBF4AxpS13Qmcn5bXBwKY0s9+A/hq2fpK4PAefa4HLhxkvAG8CbwGvJ3WXwd2LetzKnBPj+cJeAY4Pq1PSc+d1M+x/hUola1/D1hYod+7uUnrlwM3lq3vlI611SDiG3Ru/PCj0mPMUAuVWb1FxGJJN1JMoV0kaSxwOPDLiFjZx3OWpSmkWyTdDtwFXBsRTw/y8D8Drpa0EzAX+B1wSxQjr/58l+IX8vuB04HfRMS9Zdt3Bj4u6bUez1uLYrRQkaSjgK8Dk4G1gTEMbUZiNjBH0sYR0UkxurmvLD/VxDfU3Jh14yk1W92cB+whaRvgC8BGpJMF+hIRRwEfB26jmH57QtIxgzloRNwCfAA4DRgHXAzcLmn0AE/tiIj5EXE3sD/wfUmfKts+iuKX9PY9HltRjC56kXQgcA7F1Nm+wA7AD4Gxg3lNya3Ai8BXUgE/mKIIVR1fDbkx68YjHFvdlJ88sA1lJwv0JyKeAJ4AzpD0C+Bo4Nw+uv+V4sSBnvt4GbgMuEzSBcAfgG2Bx6sJPCJelHQOcLakHSIigAcoRmmLovoz5T4FPBwR734wL2lyNa+hQkyrJF0CHAosoDjb7/KyLlXFV2tuzMAjHFvNxHsnD3wN+Bx9nCzQJZ1hNTOdqba5pN0pPvx+qp+nPQt8RtKmkjZK+zlN0pckbZVOVJhO8dnMnwb5Ev4L2JpiJNG1Phq4XtInJU1OsZ4m6e/62MfTwEclTZP0IUnHUXy/pudreL+k3SVtJOlv+onpIooP/X8A3JCKR3m8/cZXx9xYi3PBsdXRLGAdoAP47QB9X6f4rOFy4P+Aq4H/pfh+TF++TTEFtxD4c2p7k2La6kGKv/o/BuwTESsGE3hEdFD8gv+hpDFpfXeKaa1rKIrJJcDmwJI+dnMu8GvgAuBhYFd6T79dB/wGuDG9hhP6iekx4BGKqbKLKsQ7UHx1yY2ZilG/mZlZY3mEY2ZmWbjgmJlZFi44ZmaWxbCcFr1ixQp/cGRmNoKNHz9ePds8wjEzsyxccMzMLIumLjilUmm4Q1itOB/dOR+9OSfdOR/dNTofVRUcSQslPS7pEUkPpLYNJN0mqZR+rp/aJelsSfMlPSZpx0a+ADMzaw6DGeF8JiK2j4id0vpJwNyIaKe4+N9JqX0fim9+t1Ncz+rn9QrWzMyaVy1TatN476qzsymulNvVflEU7gEmSNqk0g7MzKx1VHVpG0nPAssobtx0bkScJ2l5RExI2wUsi4gJkm4ATk+Xa0fSXODEiHiga3/lp0V7DtXMbGRob3/vFk+VTouu9ns4e6SbY20M3Cbpj+UbI+Lde8PXEuBglUqlmp4/0jgf3TkfvTkn3Tkf3TU6H1UVnIhYnH52SroW2AXokLRJRCxJU2adqftiYLOyp09KbXV3zdLRtL3zeq/2w7dauxGHMzOzGgz4GY6ktSWt27VMcY+SJ4A5wIzUbQbFbXZJ7Yels9V2A1ZERF+XYTczsxZRzQinDbi2+JiGMcClEfE7SfcDV0o6EngOOCj1v4nitrjzgTeAI+oetZmZNZ0BC05ELAC2q9D+ErBnhfYAjq1LdGZmNmI09ZUGzMysebjgmJlZFi44ZmaWhQuOmZll4YJjZmZZuOCYmVkWLjhmZpaFC46ZmWXhgmNmZlm44JiZWRYuOGZmloULjpmZZeGCY2ZmWbjgmJlZFi44ZmaWhQuOmZll4YJjZmZZuOCYmVkWLjhmZpaFC46ZmWXhgmNmZlm44JiZWRYuOGZmlkXVBUfSaEkPS7ohrW8h6V5J8yVdIWmN1L5mWp+ftk9uTOhmZtZMBjPCOQ6YV7Y+EzgzIrYElgFHpvYjgWWp/czUz8zMWlxVBUfSJODzwPlpXcBU4KrUZTawf1qeltZJ2/dM/c3MrIWNqbLfT4ETgHXT+obA8ohYmdYXARPT8kTgeYCIWClpRer/YqUdl0qlIYTdZTQdnR299zlqVQ37bG615XPkcT56c066cz66qyUf7e3t/W4fsOBI2g/ojIgHJU0ZciR9GCjAfi1dQNvGbRX2uXYNETWvUqlUWz5HGOejN+ekO+eju0bno5oRzieAL0jaFxgHrAecBUyQNCaNciYBi1P/xcBmwCJJY4DxwEt1j9zMzJrKgJ/hRMTJETEpIiYDBwO3R8R04A7ggNRtBnB9Wp6T1knbb4+IqGvUZmbWdGr5Hs6JwLckzaf4jGZWap8FbJjavwWcVFuIZmY2ElR70gAAEXEncGdaXgDsUqHPm8CBdYjNzMxGEF9pwMzMsnDBMTOzLFxwzMwsCxccMzPLwgXHzMyycMExM7MsXHDMzCwLFxwzM8vCBcfMzLJwwTEzsyxccMzMLAsXHDMzy8IFx8zMsnDBMTOzLFxwzMwsCxccMzPLwgXHzMyycMExM7MsXHDMzCwLFxwzM8vCBcfMzLJwwTEzsywGLDiSxkm6T9Kjkp6U9IPUvoWkeyXNl3SFpDVS+5ppfX7aPrmxL8HMzJpBNSOct4CpEbEdsD2wt6TdgJnAmRGxJbAMODL1PxJYltrPTP3MzKzFDVhwovBaWh2bHgFMBa5K7bOB/dPytLRO2r6nJNUtYjMza0pjqukkaTTwILAlcA7wDLA8IlamLouAiWl5IvA8QESslLQC2BB4sdK+S6XSkIOH0XR0dvTe56hVNeyzudWWz5HH+ejNOenO+eiulny0t7f3u72qghMRq4DtJU0ArgW2HnJEPQwUYL+WLqBt47YK+1y7hoiaV6lUqi2fI4zz0Ztz0p3z0V2j8zGos9QiYjlwB7A7MEFSV8GaBCxOy4uBzQDS9vHAS3WJ1szMmlY1Z6m9L41skLQW8FlgHkXhOSB1mwFcn5bnpHXS9tsjIuoZtJmZNZ9qptQ2AWanz3FGAVdGxA2SngIul/Rj4GFgVuo/C/i1pPnAy8DBDYjbzMyazIAFJyIeA3ao0L4A2KVC+5vAgXWJzszMRgxfacDMzLJwwTEzsyxccMzMLAsXHDMzy8IFx8zMsnDBMTOzLFxwzMwsCxccMzPLwgXHzMyycMExM7MsXHDMzCwLFxwzM8vCBcfMzLJwwTEzsyxccMzMLAsXHDMzy8IFx8zMsnDBMTOzLFxwzMwsCxccMzPLwgXHzMyycMExM7MsBiw4kjaTdIekpyQ9Kem41L6BpNskldLP9VO7JJ0tab6kxyTt2OgXYWZmq79qRjgrgW9HxLbAbsCxkrYFTgLmRkQ7MDetA+wDtKfH0cDP6x61mZk1nQELTkQsiYiH0vKrwDxgIjANmJ26zQb2T8vTgIuicA8wQdImdY/czMyaypjBdJY0GdgBuBdoi4gladNSoC0tTwSeL3vaotS2hApKpdJgQuhhNB2dHb33OWpVDftsbrXlc+RxPnpzTrpzPrqrJR/t7e39bq+64EhaB7gaOD4iXpH07raICEnRiAD7tXQBbRu39Wpub1976PtsYqVSqbZ8jjDOR2/OSXfOR3eNzkdVZ6lJGktRbC6JiGtSc0fXVFn62ZnaFwOblT19UmozM7MWVs1ZagJmAfMi4oyyTXOAGWl5BnB9Wfth6Wy13YAVZVNvZmbWoqqZUvsEcCjwuKRHUtspwOnAlZKOBJ4DDkrbbgL2BeYDbwBH1DViMzNrSgMWnIi4G1Afm/es0D+AY2uMy8zMRphBnaU2Elz49Ot9bjt8q9Y82cDMLAdf2sbMzLJwwTEzsyxccMzMLAsXHDMzy8IFx8zMsnDBMTOzLFxwzMwsCxccMzPLwgXHzMyycMExM7MsRuSlbfq7fI2ZmQ0Pj3DMzCwLFxwzM8vCBcfMzLJwwTEzsyxccMzMLAsXHDMzy8IFx8zMshiR38MZKt9+2syscTzCMTOzLFxwzMwsCxccMzPLYsDPcCT9CtgP6IyIj6S2DYArgMnAQuCgiFgmScBZwL7AG8DhEfFQY0LPa6jXZ/NnP2ZmhWpGOBcCe/doOwmYGxHtwNy0DrAP0J4eRwM/r0+YZmbW7AYsOBHxe+DlHs3TgNlpeTawf1n7RVG4B5ggaZN6BWtmZs1rqKdFt0XEkrS8FGhLyxOB58v6LUptS+hDqVQaYggAo+no7Kjh+Y1XGrUq7/FqyufI43z05px053x0V0s+2tvb+91e8/dwIiIkxVCfP1CA/Vq6gLaN2wbuN4za2/N9hlMqlWrL5wjjfPTmnHTnfHTX6HwM9Sy1jq6psvSzM7UvBjYr6zcptZmZWYsbasGZA8xIyzOA68vaD1NhN2BF2dSbmZm1sGpOi74MmAJsJGkR8H3gdOBKSUcCzwEHpe43UZwSPZ/itOgjGhCzmZk1oQELTkQc0semPSv0DeDYWoMyM7ORx1caMDOzLFxwzMwsCxccMzPLwgXHzMyy8A3YGqyvi376op5m1mo8wjEzsyw8whlBrlk6mrZ3PKIys9WTRzhmZpaFC46ZmWXhgmNmZlm44JiZWRYuOGZmloULjpmZZeGCY2ZmWbjgmJlZFv7i5zDp65I3A/EXOM2sWXmEY2ZmWbjgmJlZFp5Ss7rrb7rQU4JmrcsFx0bsLRR8MVOz1Yun1MzMLAuPcJrMUM9uMzMbbg0pOJL2Bs4CRgPnR8TpjTiOVc+FysyGW90LjqTRwDnAZ4FFwP2S5kTEU/U+lg0fFzAzGyxFRH13KO0OnBoRe6X1kwEi4t+7+qxYsaK+BzUzs9XK+PHj1bOtEScNTASeL1tflNrMzKyF+Sw1MzPLohEnDSwGNitbn5Ta3lVpqGVmZiNbI0Y49wPtkraQtAZwMDCnAccxM7MmUvcRTkSslPRN4BaK06J/FRFP1vs4ZmbWXBryGU5E3BQRH46ID0XEafXev6S9JT0tab6kk+q9/2Yg6VeSOiU9Uda2gaTbJJXSz/WHM8acJG0m6Q5JT0l6UtJxqb0lcyJpnKT7JD2a8vGD1L6FpHvTe+eKNAvRMiSNlvSwpBvSeqvnY6GkxyU9IumB1Naw90zTnTRQ9j2ffYBtgUMkbTu8UQ2LC4G9e7SdBMyNiHZgblpvFSuBb0fEtsBuwLHp/0Wr5uQtYGpEbAdsD+wtaTdgJnBmRGwJLAOOHMYYh8NxwLyy9VbPB8BnImL7iNgprTfsPdN0BQfYBZgfEQsi4q/A5cC0YY4pu4j4PfByj+ZpwOy0PBvYP2tQwygilkTEQ2n5VYpfKhNp0ZxE4bW0OjY9ApgKXJXaWyYfAJImAZ8Hzk/rooXz0Y+GvWeaseD4ez59a4uIJWl5KdA2nMEMF0mTgR2Ae2nhnKTpo0eATuA24BlgeUSsTF1a7b3zU+AE4J20viGtnQ8o/gi5VdKDko5ObQ17z/jinSNURISklruig6R1gKuB4yPileKP2EKr5SQiVgHbS5oAXAtsPcwhDRtJ+wGdEfGgpCnDHc9qZI+IWCxpY+A2SX8s31jv90wzjnAG/J5PC+uQtAlA+tk5zPFkJWksRbG5JCKuSc0tnROAiFgO3AHsDkyQ1PWHZiu9dz4BfEHSQopp+KkUFxhu1XwAEBGL089Oij9KdqGB75lmLDj+nk/f5gAz0vIM4PphjCWrNB8/C5gXEWeUbWrJnEh6XxrZIGktiovpzqMoPAekbi2Tj4g4OSImRcRkit8Zt0fEdFo0HwCS1pa0btcy8DngCRr4nqn7xTtzkLQvxXxs1/d86n7q9epO0mXAFGAjoAP4PnAdcCXwAeA54KCI6HliwYgkaQ/gv4HHeW+O/hSKz3FaLieSPkbxge9oij8sr4yIH0r6IMVf+BsADwNfjYi3hi/S/NKU2nciYr9Wzkd67dem1THApRFxmqQNadB7pikLjpmZNZ9mnFIzM7Mm5IJjZmZZuOCYmVkWLjhmZpaFC46ZmWXhgmNmZlm44NiIIulmSTOq6Pda+h7CsJA0WVKUfct9sM8/RdL59Y7LrJH8PRxreZIuBBZFxPcyHnMy8CwwtuzikX31nQJcHBGTGh+ZWeN4hGPWAEMduZiNZC441nQknSjpqh5tZ0k6W9Kdkr6e2raUdJekFZJelHRFWf9I248GpgMnpGm235YdY7GkV1XcXXbPAWI6VdJVki6W9ApwuKRRkk6S9IyklyRdKWmDPp5/hKR56XgLJB2T2tcGbgY2TfG9JmnTdLyLU5+bVdzWvXx/j0r6UlreOt258eX0Wg4q67evirukvppe73eq/GcwGzQXHGtGlwP7ll14cDRwEHBpj34/Am4F1qe4EvB/9txRRJwHXAL8R0SsExH/IGkr4JvAzhGxLrAXsLCKuKZR3MxrQtrnP1PcvOrTwKYUd5Q8p4/ndgL7AesBRwBnStoxIl6nuLvtCym+dSLihR7PvQw4pGsl3el0c+DGVLBuo8jNxhQXrvxZ2V1yZwHHpNf5EeD2Kl6n2ZC44FjTiYjngIeAL6amqcAbEXFPj65vU/zi3TQi3oyIu6s8xCpgTWBbSWMjYmFEPFPF8/4QEddFxDsR8RfgG8B3I2JRuiDkqcABlabbIuLGiHgm3anzLopC+ckq472W4r43m6f16cA16Zj7AQsj4oKIWBkRD1PcwuHA1Pft9DrXi4hlXXdNNWsEFxxrVpfy3l/1X6H36AaKuzsKuE/Sk5K+Vs2OI2I+cDxFgeiUdLmkTat46vM91jcHrpW0XNJyitsDrKLCHRQl7SPpnjTttRzYl+JK4NXE+ypwI8XoBYq8XFIWw65dMaR9Twfen7Z/OR3ruTT9uHs1xzQbChcca1a/AaaouE/9F6lQcCJiaUQcFRGbAsdQTCVtWWFfvU7VjIhLI2IPil/YAcysIqae+3ke2CciJpQ9xnXd9KqLpDUpRh0/obi97wTgJopiWTG+Ci4DDkkFYxzFfV66YrirRwzrRMQ/pdd5f0RMo5hu67q9hVlDuOBYU4qIPwN3AhcAz0bEvJ59JB2YChIUn58E790rp1wH8MGy520laWoqBG8Cf+njeQP5BXBa11RXuinatAr91qCYwvszsFLSPhQ3wyqPb0NJ4/s51k0UxfGHwBUR0RXvDcCHJR0qaWx67CxpG0lrSJouaXxEvA28MsTXaVYVFxxrZpcCf0/l6TSAnYF7Jb1GcRfD4yJiQYV+syg+x1gu6TqKX/6nAy8CSyn++j95CPGdlY57q6RXgXuAXXt2SlNi/0IxulhGMUU4p2z7HylGMAtSjL2m99LnNdfQIx9p35+jmG57Ib2emek1AhwKLExn1n2DYrrNrCH8xU8zM8vCIxwzM8vCBcesSukLlq9VeJwy3LGZNQNPqZmZWRYe4ZiZWRYuOGZmloULjpmZZeGCY2ZmWfw/7L3hnIVENHsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693c62a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693a06ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687edf3c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687ee30f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687d6ee10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687cd5f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Feature Exploration\n",
    "\n",
    "display(Markdown('## String Features'))\n",
    "for col in df.select_dtypes(include=['object']):\n",
    "    graph.figure(figsize=(16, 3))\n",
    "    graph.title(col.replace('_', ' ').title())\n",
    "    sns.countplot(df[col])\n",
    "    graph.show()\n",
    "    \n",
    "display(Markdown('## Numeric Features'))\n",
    "for col in df.select_dtypes(include=['int64', 'float64']):\n",
    "    graph.figure(figsize=(6, 3))\n",
    "    graph.title(col.replace('_', ' ').title())\n",
    "    sns.distplot(df[col], kde=False)\n",
    "    graph.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f57147a0978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Target Variable\n",
    "sns.countplot(df['success'])\n",
    "graph.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This dataset has very interesting biases. I'm going to need a score that deals with this well. Something between AUC, F1 Score, Matthews Correlation Coeffiecent.\n",
    "\n",
    "I'm currently leaning on the Matthew and F1 score.\n",
    "\n",
    "Also looks like Stanford did an absolutely amazing job cleaning this data! So it will be dope!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preping Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1472, 19)\n"
     ]
    }
   ],
   "source": [
    "# This feature order allows for easier slicing so don't fuck with it.\n",
    "features = [\n",
    "    'success',\n",
    "    'age',\n",
    "    'partner_age',\n",
    "    'age_diff_abs',\n",
    "    'children',\n",
    "    'visits_relatives',\n",
    "    'education',\n",
    "    'marital_status',\n",
    "    'partner_education',\n",
    "    'gender',\n",
    "    'house',\n",
    "    'income',\n",
    "    'msa',\n",
    "    'rent',\n",
    "    'political',\n",
    "    'religion',\n",
    "    'work',\n",
    "    'gender_older',\n",
    "    'education_difference'\n",
    "]\n",
    "\n",
    "df = df[features]\n",
    "df = df[df['house'] != 'boat, rv, van, etc.']  # perfect information\n",
    "\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['age', 'partner_age', 'age_diff_abs', 'children', 'visits_relatives', 'education', 'marital_status', 'partner_education', 'gender', 'house', 'income', 'msa', 'rent', 'political', 'religion', 'work', 'gender_older', 'education_difference'] [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17]\n",
      "(1472,) (1472, 18)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[ 52.,  48.,   4., ...,   5.,   1.,   1.],\n",
       "       [ 28.,  30.,   2., ...,   5.,   0.,   0.],\n",
       "       [ 31.,  40.,   9., ...,   5.,   1.,   1.],\n",
       "       ..., \n",
       "       [ 33.,  33.,   0., ...,   5.,   0.,   1.],\n",
       "       [ 29.,  29.,   0., ...,   5.,   0.,   0.],\n",
       "       [ 37.,  38.,   1., ...,   5.,   1.,   0.]])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1272, 18) (200, 18) (1272,) (200,)\n"
     ]
    }
   ],
   "source": [
    "# Splitting and preping data for Machine Learning\n",
    "feature_names = df.columns[1:].tolist()\n",
    "feature_indices = np.arange(len(feature_names))\n",
    "print(feature_names, feature_indices)\n",
    "\n",
    "def split_for(obj, continuous=True):\n",
    "    return obj[:5] if continuous else obj[5:]\n",
    "\n",
    "y, x = df['success'].values, df[df.columns[1:]].values\n",
    "\n",
    "# Encode Labels\n",
    "label_encoders = {}\n",
    "for col_index, feature_name in zip(split_for(feature_indices, False), split_for(feature_names, False)):\n",
    "    le = LabelEncoder()\n",
    "    x[:, col_index] = le.fit_transform(x[:, col_index])\n",
    "    label_encoders[feature_name] = le\n",
    "    \n",
    "# Unify datatypes for models like Neural Networks\n",
    "x = x.astype(float)\n",
    "\n",
    "# How does it look?\n",
    "print(y.shape, x.shape)\n",
    "display(x)\n",
    "\n",
    "# Train test stratified split (This should get us pretty good resolution)\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=200, stratify=y, random_state=42)\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a feature preprocessing pipeline since you cannot* standardise One Hot Features\n",
    "numeric_branch = Pipeline([\n",
    "    ('get-numeric', GetSubsetTransform(split_for(feature_indices, continuous=True))),\n",
    "    ('standardise', StandardScaler())\n",
    "])\n",
    "categorical_branch = Pipeline([\n",
    "    ('get-categorical', GetSubsetTransform(split_for(feature_indices, continuous=False))),\n",
    "    ('one-hot', OneHotEncoder(sparse=False))\n",
    "])\n",
    "preprocessor = FeatureUnion(transformer_list=[\n",
    "    ('numeric-branch', numeric_branch),\n",
    "    ('categorical-branch', categorical_branch)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, xs, ys):\n",
    "    prediction = model.predict(xs)\n",
    "    print(f'F1       = {f1_score(prediction, ys)}')\n",
    "    print(f'ROC AUC  = {roc_auc_score(prediction, ys)}')\n",
    "    print(f'Matthews = {matthews_corrcoef(prediction, ys)}')\n",
    "    print(f'Accuracy = {accuracy_score(prediction, ys)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Random Forest\n",
    "forest_grid_search = GridSearchCV(\n",
    "    RandomForestClassifier(class_weight='balanced'),\n",
    "    param_grid={\n",
    "        'n_estimators': [50, 100, 400],\n",
    "        'max_features': ['sqrt', 'log2', None, 10, 15, 40],\n",
    "        'min_samples_split': [5, 15, 25],\n",
    "        'max_depth': [15, 20, 30]\n",
    "    },\n",
    "    cv=RepeatedStratifiedKFold(n_repeats=10, n_splits=3),\n",
    "    n_jobs=-1,\n",
    "    scoring='roc_auc'\n",
    ")\n",
    "forest = make_pipeline(preprocessor, forest_grid_search)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#### Forest (Training) ####\n",
      "F1       = 0.8954570333880679\n",
      "ROC AUC  = 0.7927878449958182\n",
      "Matthews = 0.6532665744438857\n",
      "Accuracy = 0.8498427672955975\n",
      "None\n",
      "{'max_depth': 20, 'max_features': 15, 'min_samples_split': 25, 'n_estimators': 400}\n",
      "CPU times: user 12 s, sys: 1.93 s, total: 13.9 s\n",
      "Wall time: 1min 27s\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687d34f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "# Train Forest\n",
    "forest.fit(x_train, y_train)\n",
    "\n",
    "print('#### Forest (Training) ####')\n",
    "print(evaluate(forest, x_train, y_train))\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(forest.predict_proba(x_train)[:, 1], y_train)\n",
    "print(forest.named_steps['gridsearchcv'].best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Neural Network\n",
    "def nnet():\n",
    "    # Currently creates 78 features (checked using `forest.named_steps['gridsearchcv'].best_estimator_.n_features_`)\n",
    "    # Must use a sequential model for Neural Networks to work in SKLearn\n",
    "    model = Sequential()\n",
    "    model.add(Dense(35, input_dim=78, activation='elu'))\n",
    "    model.add(Dropout(0.1))\n",
    "    model.add(Dense(35, activation='elu'))\n",
    "    model.add(Dropout(0.1))\n",
    "    model.add(Dense(35, activation='elu'))\n",
    "    model.add(Dropout(0.1))\n",
    "    model.add(Dense(1, activation='sigmoid'))\n",
    "    \n",
    "    model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    return model\n",
    "sk_net_model = KerasClassifier(\n",
    "    nnet, \n",
    "    validation_split=0.2, epochs=int(1e6), batch_size=500, \n",
    "    callbacks=[EarlyStopping(monitor='val_loss', patience=15)], verbose=2\n",
    ")\n",
    "sk_net = make_pipeline(preprocessor, sk_net_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1017 samples, validate on 255 samples\n",
      "Epoch 1/1000000\n",
      " - 0s - loss: 0.5658 - acc: 0.7119 - val_loss: 0.4760 - val_acc: 0.8078\n",
      "Epoch 2/1000000\n",
      " - 0s - loss: 0.5219 - acc: 0.7670 - val_loss: 0.4531 - val_acc: 0.8000\n",
      "Epoch 3/1000000\n",
      " - 0s - loss: 0.5205 - acc: 0.7650 - val_loss: 0.4475 - val_acc: 0.8000\n",
      "Epoch 4/1000000\n",
      " - 0s - loss: 0.5125 - acc: 0.7620 - val_loss: 0.4443 - val_acc: 0.7922\n",
      "Epoch 5/1000000\n",
      " - 0s - loss: 0.5026 - acc: 0.7689 - val_loss: 0.4393 - val_acc: 0.7961\n",
      "Epoch 6/1000000\n",
      " - 0s - loss: 0.4987 - acc: 0.7709 - val_loss: 0.4340 - val_acc: 0.8000\n",
      "Epoch 7/1000000\n",
      " - 0s - loss: 0.4841 - acc: 0.7778 - val_loss: 0.4297 - val_acc: 0.8118\n",
      "Epoch 8/1000000\n",
      " - 0s - loss: 0.4808 - acc: 0.7827 - val_loss: 0.4276 - val_acc: 0.8118\n",
      "Epoch 9/1000000\n",
      " - 0s - loss: 0.4727 - acc: 0.7847 - val_loss: 0.4283 - val_acc: 0.8078\n",
      "Epoch 10/1000000\n",
      " - 0s - loss: 0.4742 - acc: 0.7729 - val_loss: 0.4282 - val_acc: 0.8000\n",
      "Epoch 11/1000000\n",
      " - 0s - loss: 0.4602 - acc: 0.7797 - val_loss: 0.4296 - val_acc: 0.8000\n",
      "Epoch 12/1000000\n",
      " - 0s - loss: 0.4671 - acc: 0.7807 - val_loss: 0.4296 - val_acc: 0.7922\n",
      "Epoch 13/1000000\n",
      " - 0s - loss: 0.4499 - acc: 0.7837 - val_loss: 0.4243 - val_acc: 0.7922\n",
      "Epoch 14/1000000\n",
      " - 0s - loss: 0.4494 - acc: 0.7866 - val_loss: 0.4192 - val_acc: 0.8118\n",
      "Epoch 15/1000000\n",
      " - 0s - loss: 0.4440 - acc: 0.7866 - val_loss: 0.4158 - val_acc: 0.8078\n",
      "Epoch 16/1000000\n",
      " - 0s - loss: 0.4496 - acc: 0.7827 - val_loss: 0.4128 - val_acc: 0.8078\n",
      "Epoch 17/1000000\n",
      " - 0s - loss: 0.4455 - acc: 0.7935 - val_loss: 0.4095 - val_acc: 0.8118\n",
      "Epoch 18/1000000\n",
      " - 0s - loss: 0.4469 - acc: 0.7886 - val_loss: 0.4055 - val_acc: 0.8118\n",
      "Epoch 19/1000000\n",
      " - 0s - loss: 0.4371 - acc: 0.7876 - val_loss: 0.4038 - val_acc: 0.8039\n",
      "Epoch 20/1000000\n",
      " - 0s - loss: 0.4236 - acc: 0.7925 - val_loss: 0.4044 - val_acc: 0.7843\n",
      "Epoch 21/1000000\n",
      " - 0s - loss: 0.4124 - acc: 0.8112 - val_loss: 0.4046 - val_acc: 0.8000\n",
      "Epoch 22/1000000\n",
      " - 0s - loss: 0.4245 - acc: 0.8112 - val_loss: 0.4048 - val_acc: 0.8039\n",
      "Epoch 23/1000000\n",
      " - 0s - loss: 0.4150 - acc: 0.8083 - val_loss: 0.4042 - val_acc: 0.8000\n",
      "Epoch 24/1000000\n",
      " - 0s - loss: 0.4096 - acc: 0.8151 - val_loss: 0.4033 - val_acc: 0.8078\n",
      "Epoch 25/1000000\n",
      " - 0s - loss: 0.4076 - acc: 0.8171 - val_loss: 0.4023 - val_acc: 0.8039\n",
      "Epoch 26/1000000\n",
      " - 0s - loss: 0.4070 - acc: 0.8083 - val_loss: 0.4018 - val_acc: 0.7843\n",
      "Epoch 27/1000000\n",
      " - 0s - loss: 0.4068 - acc: 0.8151 - val_loss: 0.4008 - val_acc: 0.7961\n",
      "Epoch 28/1000000\n",
      " - 0s - loss: 0.4024 - acc: 0.8122 - val_loss: 0.3981 - val_acc: 0.8000\n",
      "Epoch 29/1000000\n",
      " - 0s - loss: 0.4020 - acc: 0.8181 - val_loss: 0.3966 - val_acc: 0.8000\n",
      "Epoch 30/1000000\n",
      " - 0s - loss: 0.3928 - acc: 0.8250 - val_loss: 0.3949 - val_acc: 0.8078\n",
      "Epoch 31/1000000\n",
      " - 0s - loss: 0.3913 - acc: 0.8269 - val_loss: 0.3933 - val_acc: 0.8039\n",
      "Epoch 32/1000000\n",
      " - 0s - loss: 0.4036 - acc: 0.8112 - val_loss: 0.3925 - val_acc: 0.8000\n",
      "Epoch 33/1000000\n",
      " - 0s - loss: 0.3888 - acc: 0.8092 - val_loss: 0.3912 - val_acc: 0.8039\n",
      "Epoch 34/1000000\n",
      " - 0s - loss: 0.3874 - acc: 0.8250 - val_loss: 0.3899 - val_acc: 0.8039\n",
      "Epoch 35/1000000\n",
      " - 0s - loss: 0.3865 - acc: 0.8230 - val_loss: 0.3892 - val_acc: 0.8039\n",
      "Epoch 36/1000000\n",
      " - 0s - loss: 0.3840 - acc: 0.8250 - val_loss: 0.3885 - val_acc: 0.8118\n",
      "Epoch 37/1000000\n",
      " - 0s - loss: 0.3878 - acc: 0.8181 - val_loss: 0.3866 - val_acc: 0.8078\n",
      "Epoch 38/1000000\n",
      " - 0s - loss: 0.3830 - acc: 0.8250 - val_loss: 0.3858 - val_acc: 0.8118\n",
      "Epoch 39/1000000\n",
      " - 0s - loss: 0.3818 - acc: 0.8279 - val_loss: 0.3888 - val_acc: 0.8196\n",
      "Epoch 40/1000000\n",
      " - 0s - loss: 0.3769 - acc: 0.8230 - val_loss: 0.3953 - val_acc: 0.8118\n",
      "Epoch 41/1000000\n",
      " - 0s - loss: 0.3858 - acc: 0.8240 - val_loss: 0.3962 - val_acc: 0.8118\n",
      "Epoch 42/1000000\n",
      " - 0s - loss: 0.3952 - acc: 0.8338 - val_loss: 0.3934 - val_acc: 0.8078\n",
      "Epoch 43/1000000\n",
      " - 0s - loss: 0.3780 - acc: 0.8220 - val_loss: 0.3888 - val_acc: 0.8078\n",
      "Epoch 44/1000000\n",
      " - 0s - loss: 0.3724 - acc: 0.8299 - val_loss: 0.3892 - val_acc: 0.8078\n",
      "Epoch 45/1000000\n",
      " - 0s - loss: 0.3793 - acc: 0.8230 - val_loss: 0.3904 - val_acc: 0.8118\n",
      "Epoch 46/1000000\n",
      " - 0s - loss: 0.3855 - acc: 0.8250 - val_loss: 0.3907 - val_acc: 0.8157\n",
      "Epoch 47/1000000\n",
      " - 0s - loss: 0.3930 - acc: 0.8151 - val_loss: 0.3923 - val_acc: 0.8157\n",
      "Epoch 48/1000000\n",
      " - 0s - loss: 0.3792 - acc: 0.8299 - val_loss: 0.3948 - val_acc: 0.8157\n",
      "Epoch 49/1000000\n",
      " - 0s - loss: 0.3742 - acc: 0.8201 - val_loss: 0.3980 - val_acc: 0.8118\n",
      "Epoch 50/1000000\n",
      " - 0s - loss: 0.3771 - acc: 0.8319 - val_loss: 0.4007 - val_acc: 0.8000\n",
      "Epoch 51/1000000\n",
      " - 0s - loss: 0.3721 - acc: 0.8387 - val_loss: 0.4023 - val_acc: 0.8000\n",
      "Epoch 52/1000000\n",
      " - 0s - loss: 0.3722 - acc: 0.8289 - val_loss: 0.4049 - val_acc: 0.8039\n",
      "Epoch 53/1000000\n",
      " - 0s - loss: 0.3699 - acc: 0.8250 - val_loss: 0.4072 - val_acc: 0.7961\n",
      "#### Neural Network ####\n",
      "F1       = 0.8878741755454085\n",
      "ROC AUC  = 0.7566876076603506\n",
      "Matthews = 0.5026806851200609\n",
      "Accuracy = 0.8262578616352201\n",
      "None\n",
      "CPU times: user 3.16 s, sys: 516 ms, total: 3.68 s\n",
      "Wall time: 2.02 s\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbIAAAF3CAYAAADAVs4bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzs3Xd4VNXWwOHfnj4pJPQOUsWCCqIgIoJYsWLXz16uAhaKCggqCgqodMVyLddeLoiCYkGuil1QsYBKkA6hQ0iZPvv7I4A5TMpAZubMTNb7PD5m1j4zs3JIZuWcs8/aSmuNEEIIkaosZicghBBCVIcUMiGEEClNCpkQQoiUJoVMCCFESpNCJoQQIqVJIRNCCJHSElLIlFIvKKW2KKV+r2BcKaWmKaVWKKV+VUp1TkReQgghUl+ijsj+A5xZyfhZQLs9//0LeCoBOQkhhEgDCSlkWuuFwI5KNjkfeFmX+g7IVUo1TkRuQgghUpvN7AT2aAqsK/N4/Z5YftmNCgoKpA2JEEKksZycHHWgz5HJHkIIIVJashSyDUDzMo+b7YkJIYQQlUqWU4tzgNuUUm8CXYECrXV+Fc8RVcjLy6Ndu3Zmp5ESZF9FT/ZV9BK9r5buCPDhOi+eYBh/GKb/XkQDt4UtnnDCctjf9Ydm0Dzrn1JzzG/zOfvTJ7GGQ/tiRS99Xq33SEghU0q9AfQC6iml1gMPAHYArfXTwDygL7ACKAGuT0ReQgiRDnb5wny20cv1n++MGItlEetYxx4R84c0R9e1c88x2VhU6eUti4IWWdZ9jwHQGsc7L+D45JWY5bNXQgqZ1vqKKsY1MDARuQghRCr7c1eAq/+3A39IY7PA37tDVT+pHIfXthkLTRm/7wjQt4WLY+s5AGibY+OcFi6slgOeh1Eq4Mf5/KPYv/3UENZK4b+8/8G9ZhnJcmpRCCHSVn5JiG3e8o+MftzqZ3lBgAyrccrCJ+u9/LojQNta/3xMr9gdrHYuZzRz8nTPOtR2JmiKRHEh7mmjsP75iyGs7Q68t4wkdNzJ1X4LKWRCCBFju3xh1hWHeOhPB/O/qt68tYMtXofm2LikTQa+kObERg6aZ9pommnFZTvIo6qDoLbm4544DEv+WkNcZ+fgGfQI4bZHxOR9pJAJIUQVtNb8vTuIt5yzeGGtmbPGS1hrLErx7LIidgf23vKa+I/YTvXsnNXcxV1HZ1d46jARLH//gWvKvVh2G6/bhRs1xzNkPLph05i9lxQyIURK8QY1KwuD6Bi3R9jsCTHs+4J9BWmvHd4w233mzfqrzOQTcjmpcel1rMYZVjLtyXFHlXXxl7ieGYvy+wzxUPuOeO4cC1k5MX0/KWRCiKQRCGtWFAT530YfTy0twr3fabC8gupfIzLbEbUjP3Y3FIdwWhU3dcg0xIMachwWTm3qpOzBVfNMW0JPER4I+8f/xfHGDNR+f2kEup6C76Zh4HDG/D2lkAkhTPfXrgD9Pt7GxpLkPPKpjja1rHRt4GTqibnYD3bWXyoIh3C8PgPH/FkRQ/5zr8J/4Q1gic8RoxQyIURCFQbCPPdHMQ/+uJvW2VZsFsXyFDrSapJhIbecGX/Ldgbp1cRJ94alp/pa17JxRGADhx1aA24e93lwPTUW289fG8LaYsF37RCCvc6J69tLIRNCVIs3qHlnVQl/7AruOxWoNTy5tAgL0CTTati+bNFaWXhw90DtdXhu7D/C/GG4o2MW3Ro4DHGnVdEiy4o6gAkUeXmxzi75qF3bcU25F+uqvwxx7crAe9toQh2Pj3sOUsiEEFEJhjXvrvbw3Ro77/+Yz3ZfmEAUZwIP9mirXY6NLLvikeNzqLvfEVBdl4W6LmsFzxSJojasxj1pGJZtmw3xcO16eIeMJ9yibULykEImhKiU1po1RSHu/HoXX+T7KO0uF79rWfd2yuaeY2rF7fVFbFj/+BnXtFGokmJDPNSiDd7B49F16icsFylkQogKzV5VUm7/vlhpn2Nj6om51NlzxNUk00p2kkwhFxWzffUxzhceQ4WMR9vBjsfjHTga3BmJzSeh7yaESDq7/WFmr/KwrjjE478U0tBtobbTwp+7DvyUYI5DcevhWey9iuQNaY6obeeoupHNZptnWcmwSdFKKVpjf+9lnLNfjBgK9DoX3zV3gjXxZUUKmRA12KrdQTrNMl7f2OwJszmKjumDO2bRqpaN4xs4qOeyUE+uWaW3YADni49j/+rjiCHfpbcQ6Hs5mNRJRAqZEDXUwZ427JYbYtxJjehUz1H1xiI9FBfieuIBbMt+MoS13Y7v5hEEu55iUmKlpJAJUQPNXFnCTV8cWBH74rz6HF3XUbpYpBSxGkNt24Rr4nCsG1cb4jqzFp5BYwm3P8qcxMqQQiZEDbBgg5cft/rRwLurPPxRyfWvO47M4tj6DtrnlH481HZaaJQhpw1rIsuqP3FNvhdLwQ5DPNygCZ6hj6IbNTMpMyMpZEKkMU9Qc+En2/h2s7/Kba9tn8Hk7rmmdkwXycP68ze4ZjyE8nsN8VDbI/Dc+TDUyjUps0hSyIRIYSXBMOuLQqwqDDH1t0K+2ew3dLtYFuXMwyd65HJVu8yqNxQ1gv3T2ThenY7Sxkk/geN64fvXiLg0/q0OKWRCpIjd/jCbSkpbOm3zhhnw1U5Wl9PiKdritdeTPXL5PyliAiAcxvHW0zg+ejtiyN/3cvyX/CtujX+rQwqZEEmqJBim/5c7+X1HgL93V68nYVmDOmZhtyiaZ1m5sJWbLLkBWQD4vLiefQTb4oWGsFYWfFffSbDP+SYlVjUpZEKYbJs3xM49Czeu3B3ihy0+vtnsj+q61oE4qo6dt0+rKxM3RAS1e2dp49+//zDEtdOFd+BoQkd3Mymz6EghEyKGSoJhZq70sK4ohKOSAx0NTP2tiOJg7JY5buC2sMUTZnDHLC5snUHZt2+eZaVWZQmJGkvlr8U9cRiWrfmGeDi3bmnj35bJvwyNFDIhYuDDtR6uWLCj6g1jpF2OjUBYs7owxJgutbihQ2bSLHMvUoflz19wTxuFKi40xEPNWuEdMgFdt4FJmR0YKWRCHKTd/jCvryjh791B/v1HcdVPqIZr2mfwr8OyaJktTXVFbNi+/RTncxNQwYAhHjyiC97bRkNGljmJHQQpZEJEyRfSrC8K8dbKEh5dUlj1Ew5C21o2dvnDbPOGGdQxi2aZVi5unVHuisRCHBStsb//Gs6Zz0UMBXr2xXftELClVmlIrWyFSKCvNvm4deFOioNhdvoO/FpWiywrl7aufDmL4mCYo+s66NfKjdMqNyKLOAsGcb40CfvCeRFDvotuJHDuVaY1/q0OKWRC7LF0R4C7v9vFoq3+qFY+rsgDx9Zi8FHZsUtMiFjwFOOa/gC2pYsNYW2z47vxHoLdTzMpseqTQiZqHF9Is6G49L6sX7b7eeyXQpbtzAC2HPRrHlnHzpnNXRxX38EZzV0xylSI2FDbt+CaNBzr+pWGuM7MxnPHGMIdjjEps9iQQiZqhBUFAT5Z72P2qhIWbQ1U/YQqNMmw0DzLxmt96sg6XCKpWdbk4Zo0AsuubYZ4uF4jPEMnoJu0NCmz2JFCJtLWDm+IAr/m1bxiJv5aVK3XmnhCDsfVd9A4w0p9txQukRqsv3yPa8ZolNdjiIdaH4Z30MPonDomZRZbUshEWglrzcRfCnn45+rNKrztiCwuaeMm12GhZbb8mojUY/tsDs6Xp6DCxgu+wc498N46CpzpcwpcfkNF2tjlC9P+zXz8BzBRo1W2FQ14/AEeP7EeHXJttK1lQ6XgzC0hgNLGv//9N455b0QM+U+/GP8V/cGSXmcVpJCJlPTHzgB3fL0TX+if2cK/bI/u2lePRg7OP8TNpW0yyNnTtikvL492Ld3xSleIxPD7cD47Dvuizw1hrRT+K28jcPpF5uQVZ1LIRMrQWvPsH8UM+77goJ7fLsfGUyfVpkt9R4wzEyIJFO7CPWUU1hW/G8La4cTb/z5CnXuYlFj8SSETSWWbN0RRQLNgg5ddPk3Ze4T/t9HHwnzfAb1e3uWNZHKGSHtq03rck4Zh2bzBEA/n1MY7aBzh1h1MyiwxpJCJuNFa8+kGH8t2Vn7KLxCGsT/tjul792zs5N0z6mKRa10izVmW/4Z76khUkfF3KNykJZ4h49H1G5uUWeJIIRNxM+nXIsbEuECVZ0LXHLo2+Od0YX23laaZchQm0p/t+89w/vsRVGC/xr8djsF7xxjIrBkdZqSQibjY5g0x+df4NNbda+6Z9TipsTOu7yFEUtIa+7w3cb79TMRQoPvp+G68G2x2ExIzhxQyEVOeoKbP+1tYtjMYk9c7oraNU5sa73dxWBVnNHfJpA1RM4WCOF+Zhv2zORFD/guuxX/BdSnZ+Lc6pJCJgxbWmi/zfSzaGmCzJ1TlmlzntHDRulbFP3KFgTC9mrjo28KF3VKzfhGFiIqnBNeMB7H9+r0hrK1WfDfcTbDHmSYlZi4pZOKgLNnmp9fcrVFvn2VTvNCrDg5ZqkSIg6J2bMU1eQTWtSsMcZ2Rife2hwgdcaxJmZlPCpk4YL6QPqAidlx9OyM61ZIiJsRBsqz9G9fk4Vh2GH/vwnUb4h0ynnCzViZllhykkIkDsrkkxKFvbYpq2+sPzWBC11wpYEJUg/W3RbieeADlLTHEQy3b4x0yDp1b16TMkocUMlGl7d4QDyzezQ9b/CwvqHgSx8AjsigKhLnu0EyOqWuXfoVCVJPtiw9w/mdiZOPfY07A2/8+cFW+AnlNIYVMlOunrX6+2+Lnhy1+3l3tqXTb5llWfrqooUzQECJWtMYx63kcc1+NGPL3uQD/VbenXePf6pBCJoDShScfWLybhfk+CgP6gJ77Ud/6UsSEiJWAH+fzj2L/9lNDWCuF//IBBM64uMZNr6+KFLIaYoc3REnQWKB8IfhkvZffdwZ4La+kgmdW7OYOmYw9PgenXAMTIjaKduOeOgrr8l8NYW134L11FKEuPU1KLLlJIUtznqDmmv9tZ/6GA2u2W5Hj6zuY0C2Hw3LtuGxSwISIFbVlY2nj3/x1hng4OxfvoIcJtz3CpMySnxSyNPLVJh+/bg/w2QYv32/109Lp5LevNlb7dW8+LJOwhtuPzOIQWS1ZiJizrFiKa8pILIW7DPFwo+aljX8bNjUps9Qgn0ppYNnOAN3f3RIR/81/cBeDT27sZGTnbGo7LbJashBxZl28ENfTY1EBvyEean8UnjvHQlYtkzJLHVLIUtx7qz1c+9mOA3pO04x/CpxGs7EkzGlNnRxV107fFm6OlR6GQsSf1tg/nonjzRkobbx+HejWB99Nw8Auv4vRkEKWwvIKAgdUxJpmWPno7Ho0z5J/diFMFQ7heO0JHJ/Ojhjyn3sV/gtvAIvFhMRSk3yipbBftle8YOWNHTLx7N7Ftcc0RilwWBSH17ZLlw0hzObz4JoxBtuSbwxhbbHgu3YIwV7nmJRY6pJClqJCYc2jSyLX+7r76Gzu7ZSNUoq8vK20ayjrdQmRLNSu7aWNf1cvN8S1KwPvbQ8S6nicSZmlNilkKerFv4oj2kX1buJkZGe5MCxEMrKsX4Vr0nAs2zcb4uE69fEOHk+4RRuTMkt9UshSTFEgzMfrvNz1XUHEWLsc+ecUIhlZl/2Ea/p9qBLjmn2hFm3xDh6HrlPfpMzSQ8I++ZRSZwJTASvwnNZ6/H7jLYCXgNw92wzXWs9LVH7JTmvNku0BeleyfMr1h2YmMCMhRDRsX32E84XHUKGQIR48qiveAQ+AWxr/VldCCplSygo8CZwGrAcWKaXmaK2XldlsFPC21voppdThwDzgkETkl+zyCgKcMHsLwUpaIE4/MZfDatsTl5QQonJa0+iLObi+nBsxFOh9Lr6r7wSrnEWJhUTtxeOBFVrrlQBKqTeB84GyhUwDey/w5ADVb0mRJqb8VlRpEcuyKU5p6kpcQkKIygUDOJ9/jKxvPokY8l16C4G+l0vj3xhKVCFrCpRtILYe6LrfNqOBT5RStwOZwKmJSS355ReHKhy766hs+rVy0zRTlnQQIikUF+Kafj+2P342hLXdju/mewl27W1SYukrmY5rrwD+o7WeqJQ6AXhFKXWk1jpc3sZ5eXmJzc4k+V7F/za6I+L3tvVxfsMQFlUC2yFve/nPryn7KRZkX0VP9lX5HLu20frNadi25RviQXcWKy8dSHGdZiD7LkK7du2q9fxEFbINQPMyj5vtiZV1I3AmgNb6W6WUC6gHRDYRpPrfeCrY5Qtz3Ov5EfFZp9elTxSnEvPy8mrEfooF2VfRk31VPsuqP3G98iiWgp2GeLhhU3xDJtCkUTOTMkt/ieqBsghop5RqpZRyAJcDc/bbZi3QB0ApdRjgAiqeopfmfCHNGR+U/+3XdUrrGiGSifWnr3E/MiiiiBU1a0PJfU+ipYjFVUKOyLTWQaXUbcDHlE6tf0FrvVQp9RCwWGs9BxgK/FspNZjSiR/Xaa0PbKniNPHdZh9nzttW7tgpTZx0rCOzE4VIFvb57+B4bXpk49/jerGiz6W0zc41KbOaI2HXyPbcEzZvv9j9Zb5eBpyYqHyS1VNLixjxQ+TNzgCHZFuZeXpdLDLbSQjzhUM43ngKxyczI4b8fa/Af8nN6L//NiGxmieZJnvUeM//WXERA/jy/AZSxIRIBj4vrqfHYvvpK0NYKwu+a+4keMr5JiVWM0khSxIbikMM/bbiIvZR33pk2+XamBBmUwU7cE0ZiXXlH4a4drrwDhxN6OhuJmVWc0khSwI7fWGOmbmp3LFhx2Qz4IgschxSxIQwm9q4Bvek4Vi2GmcTh3Pr4R0yjnBLmc1pBilkSWBhvo9AOXfL/adXHS5oFXkPmRAi8Sx/LsE9dRSqpMgQDzVrjXfIeHTdBiZlJqSQJYGScvpPXdc+Q4qYEEnC9s18nM9NQIWMSycFj+iC9/YHwS0Nu80khSxJTegmU3aFMJ3W2Oe8gvOdFyKGAj374rt2CNjkY9Rs8i9gsvySEP2/NN5EeVkbN06rzE4UwlTBIM6XJmFfGLmalO/imwic83/S+DdJSCEz2TPLiqreSAiRWCVFuJ54ANvSHw1hbbPju2kYwROkp3kykUJmkh+3+nlyaRHvrPJEjLXPkc4dQphFbd+Ma9JwrOtXGeI6MxvPHWMJdzjapMxERaSQmeDl5cXc8fWuCsdvPVwuHAthBsvq5bgmj8Cyy7icRLh+YzxDJ6AbtzApM1EZKWQJ9u8/irj7u4pvfH771Lpkyo3PQiScdcm3uGY8iPJ5DfFQm8PwDnoEXau2SZmJqkghS5B1RUF6ztnCTl/FfZA71bPTo7EjgVkJIQBsC97D+cpU1H7LHwaPPQnvLSPBKSuwJzMpZAly/ec7Kixitx6eyaWtMziqrh2bRWZBCZEw4TCOt5/B8eFbEUP+My/Ff9ktYJHV15OdFLIECIU1i7cGyh17qXcdzj9EbnwWIuH8PlzPPoJt0ReGsFYW/P93G4HTLjQpMXGgpJDFkdaaF/4qZt5ab7njM0+ry6nN5JSFEAm3exfuqSOxrlhqCGuHC++A+wl16m5SYuJgSCGLk+JAmBPe3cLaolC549uvbYJVTiMKkXBq0zrcE4dh2bLREA/n1ME7eBzhVoealJk4WFLI4uTeHwoqLGJ1nRYpYkKYwLL8V9xTRqGKdxvioSaH4B06Hl2vkUmZieqQQhYHy3cFeGl5SYXjtx+ZlcBshBAAtu//h/PZcaig8Xp18PDOeG97EDKzTcpMVJcUshgrCoQ5/YOt5Y49f3JtOtdz0KqW7HYhEkZr7PPewPn2sxFDgR5n4Lv+LrBJN51UJp+oMbKpJMSVC7bz07byZyf+cnFDWmbL7hYioUJBnC9Pxf753IghX7/rCZx/jTT+TQPyyRojM5YWVVjELmntliImRKJ5SnA9ORrbbz8Ywtpqw3fD3QR7nGFSYiLW5NM1BrTWzFxZ/jWx2k7FMz2ltY0QiaR2bME1eQTWtX8b4jojE+/tYwgd3tmkzEQ8SCGLgQlLCtlYEo6IH1ffzvQetbHIqQshEsaydgWuScOx7NxmiIfrNcQzZAK66SHmJCbiRgpZDLyaF3k0Jh07hEg8628/4HriAZTXuDxSqNWhpY1/c+ualJmIJylkMVAUiDwaO76BNP8VIpFsn7+P86VJqPB+jX87nYi3/yhwyh+W6UoKWRy8e0ZdGmdIo1EhEiIcxjHreRzvvxYx5D/tQvxXDpTGv2lOClkcHF1XjsaESAi/D+dzE7B//z9DWCuF/4oBBM64xKTERCJJIasmT1Czy1/xGmNCiDgpKsA9dRTW5b8ZwtrhxHvLKEJdTjIpMZFoUsiqQWvNhZ9sq3pDIURMqc0bShv/bl5viIezc/EOfoRwm8NNykyYQQpZNazYHeTbzf6IuFVm2wsRN5YVS3FPuRdVWGCIhxs3L51e36CJSZkJs0ghq4aiQOQpxWPr2anlsJiQjRDpz7roC1zPPIwKGP+ADHU4Gs/tYyCrlkmZCTNJIYuxN06V+1SEiDmtsX/0No63nkZp4x+QgRNOxXfjPWCXSVY1lRSyatjqMd6vckxdOw3cMs1XiJgKBXG89gSOBe9GDPnPuxr/hTdI498aTgrZQVpbFOTST7ebnYYQ6c1bgmvGQ9h++c4Q1lYrvuuGEuzZ16TERDKRQnaQPlzrjYjZ5NKYEDGjdm4rbfy7Js8Q1+5MvLc9SOjILiZlJpKNFLKDVN5Ej16NXSZkIkT6saxfiWvicCw7thji4ToN8A4ZT7h5a5MyE8lIClmMZNoUwzvJUulCVJd16WJc0x9AeYoN8VDLdngHj0PXrmdSZiJZSSGLkVsOz8RmkQvOQlSHbeGHOP/zOCoUMsSDR3fDO+B+cGWYlJlIZlLIhBDm0xrH7BdxvPdyxFCg93n4rr4DrPJxJcp3wD8ZSqkGWustVW8phBBRCPhxvvAY9m/mRwz5LruVwFmXyfR6Uamo5tkppXKUUi8rpTzAqj2xc5VSD8Y1uySltcYflkbBQlRbcSGux++JKGLabsczcDSBvpdLERNVinbC+FOAF2gP7O0N8z1wRTySSmYfrPFw2FubmLCk0OxUhEhpams+GWMGYvtziSGus3PwDJtM6Phe5iQmUk60pxZPBZpprf1KKQ2gtd6ilGoYv9SSj9aaod/uYpMnckVoIUT0LH//gWvKvVh27zTEww2b4Rk6Ht2wmUmZiVQUbSHbDdQBNu0NKKWaA5vjkVSy2uXXFRaxVtlyIVqIaFh//BLX02NRfp8hHmp3JJ5BD0NWjkmZiVQV7afvC8B/lVL3Ahal1HHAOOCZuGWWZLxBzRkfbI2IZ9kUZ7d0cWkbmRYsRFXsn8zE8fqTkY1/u/bGd9NwcDhNykyksmgL2TjABzwPuIDXKS1ik+OUV9KZs8bD8oKgIZZhU6y/WtY+EqJK4RCO12fgmD8rYsh/9pX4L74JLNLjTRycaAtZXa31RGBi2aBSqh6Q1ksk/74jQI/3yr/b4KRGsmyEEFXyeXA9PRbbT18bwtpiwXfNYIK9zzUpMZEuoi1kK4HyVqxbTum1s7R1/6KCCsee6ZnW37oQ1aYKduCafC/WVX8a4trlxjtwNKGjupqUmUgn0RayiBs5lFJZQNpP31tbFCo3Pr5rDrlOORUiREXUxjW4Jw7Dsm2TIR6uXa+08W+LtiZlJtJNpYVMKbUK0IBbKbVyv+F6QOQJ7zTjK+fG58e75XBjh0wTshEiNVj/+BnXtPtQJUWGeKh5G7xDxqHrNDApM5GOqjoiu4nSo7E5wM1l4hrYrLVeGq/EksE7K0tYt98R2aILG9Aux25SRkIkP9vXn+B8/lFUyDg5KtjxOLwDR4Nb/ggUsVVpIdNaLwBQSjXSWu9OTErJwR/SDP52V0TcLh3uhSif1tjfexnn7BcjhgInn43vmsFgk/stRexF9VOltd6tlDoSOInSU4qqzNhDccrNVJs8IQr8xtOKrbOttMyympSREEksGMD54kTsX30UMeS75GYCZ18pPRNF3ERVyJRSNwLTgQXAacB8oA8wN36pJZ8P+tZHyS+jEEbFhbieeADbsp8MYW2z47t5OMFufUxKTNQU0R7nDwf6aq0/V0rt1Fqfq5Q6B7gwjrkllWaZVhpnyNGYEGWpbZtwTRqOdcNqQ1xn1sJz51jChx5lTmKiRol2/nhDrfXne74OK6UswAfABXHJSgiR9Cyr/sL90ICIIhau34SS+56QIiYSJtojsvVKqZZa6zVAHnA2pR09AnHLzEShsGbYdxXfCC1ETWdd8g2uJx9C+b2GeKjN4XgGPQK1ck3KTNRE0R6RTQSO3PP1WOBt4Evg4WjfSCl1plLqL6XUCqXU8Aq2uVQptUwptVQp9Xq0rx1rX23y8eE6b9UbClED1Vv8Ga4poyKKWLBLTzzDJ0sREwkX7azF58t8/b5Sqjbg1FpHddiilLICT1I6UWQ9sEgpNUdrvazMNu2AEcCJWuudSinT7phcXRjZzeOQbLk+Jmq4cBjHW0/T/KO3I4b8Z12G/9JbpPGvMMVB/dRprb2ATSk1LsqnHA+s0Fqv1Fr7gTeB8/fb5mbgSa31zj3vUX6nXpM8cryskSRqML8P15OjcexXxLSy4L1mEP7L+0sRE6ap8idPKXWtUmqyUmqAUsqmlMpRSj0GrAY6R/k+TYF1ZR6v3xMrqz3QXin1tVLqO6XUmVG+dtxd0z6Do+pKp3tRM6ndO3GPH4xt8UJDXDtceAeNJdhH5nwJc1XVa/FR4GrgG+AKoBtwAvAjcJLWekmMc2kH9AKaAQuVUh211pHtNYC8vLwYvrXR5i1W4J8F/goKdpOXl5qr1cRzP6Ub2VeRnNs30eaNaVh3GReVDWTl8Pdlt+PJrA+y3yolP1dVa9euXbWeX9U1ssuBnlrrPKXUYcBS4Aqt9VsH+D4bgOZlHjfbEytrPfC91joArFJKLae0sC0q7wWr+41XpmG4GFb8Uz9zcmrRrl3tuL1fvOTl5cV1P6UT2VeRLH/9ivvlR1HFhYa4p34TwiMm06xuQ5MySx3yc5UYVZ1azNVa5wForf8ASg6iiEFpMWqnlGqllHJQWiDn7LfNu5Qeje1dsLM9petKpq0/AAAgAElEQVSgJVx+SflLtwhRU9i+XYD70aERRSx4eGeWXzsMLUVMJJGqjsiUUqo5//RWDO73GK312qreRGsdVErdBnwMWIEXtNZLlVIPAYu11nP2jJ2ulFoGhIC7tdbbD/xbOngF/jDXfraDzzf6Evm2QiQPrbG//zrOmf+OGAr0OBPf9UMJr1qd+LyEqERVhSyT0kkdZRsMrinztaa0MFVJaz0PmLdf7P4yX2tgyJ7/THHbVzvLLWJWaa8oaoJgEOfLk7F/8UHEkO/CGwicd7U0/hVJqapCVqMW3lqYX/6R2ImNnOXGhUgbnmJcT4zG9rvxkrS22vDdeA/BE083KTEhqlbVemQ15mLRs8uKIpZtAXi0aw4XtXKbkJEQiaF2bME1cTjW9cZL0jojC+8dYwgd1smkzISIjqxyB2z1hLjn+8gmJX9e1ohG0vFepDHLmjxck0Zg2WW8vSRcrxGeoRPQTVqalJkQ0ZNCBjzzR3FErI7TQgO3dCoQ6cv6y/e4ZoxGeT2GeKhVB7yDH0Hn1DEpMyEOTI0vZN9v9vH4L4UR8cndc7HIhW2RpmyfzcX58mRUOGyIBzufiPfW+8DpMikzIQ5c1IVMKWUDjgOaaq1nKqXcAFprT+XPTG5fbfJHxHo1cXL+IXJdTKShcBjHzH/j+OCNiCH/6Rfhv2IAWOR0ukgtURUypdQRwHt7HjYCZgJ9gP+jtHVVygrqyAke17XPNCETIeLM78P57/HYf/jMENZK4b9yIIHTLzYpMSGqJ9qLQE8BY7XWbflnMc3PgZPikZSZ+h3i5gKZpSjSTeEu3I8OjSxiDifeO8ZIERMpLdpTix2Bl/Z8rQG01kVKqYy4ZGWitjk1/rKhSDNq83rcE4dj2bzeEA/Xqo138DjCrTuYlJkQsRHtp/YaoBPw096AUqoL8Hc8kkqUUFjz7LLIGYtCpAtL3u+4p9yLKtptiIebtMQzZDy6fmOTMhMidqItZPcDHyilZgAOpdTdwECgf9wyi7MFG7xc9ElCWzkKkVDWHz7H9ezDqEDAEA92OAbvHWMgM9ukzISIragKmdZ6jlIqn9JVnL8GDgUu01p/H8/k4mXJNn+FRayWXabcixSnNfYP38L51tMRQ4Hup+G74W6wy0KxIn1EO2uxttZ6ERWsDZZKtnpC9J67tcLxc2XavUhloSDOV6Zh/2z/VZLAf/61+PtdJ41/RdqJ9tTiBqXUp8BrwJxUvnds8VY/kRPuoWMdO7NOr0sDt9xDI1KUtwTXjIew/fKdIaytVnzX30XwpLNMSkyI+Ip2+n0r4FNgMLBJKfWKUuospVTKfeoHy6liQ4/K4svzG0gREylL7dyG+5E7I4uYOxPv0EeliIm0FlUh01pv1lpP01p3o3T24l/A48DGeCaXCA3cFkZ1rmV2GkIcNMu6lbgf6o91TZ4hHq7bEM+o6YSOONakzIRIjIO5aSpnz3/ZQMrPXT+uvgMl1wxEirL+vhjX9PtR3hJDPNSyPd4h49C5dU3KTIjEiXayR3tKW1FdSWkR+y9wudb6mzjmFnPFgTA3fr7D7DSEiAnbwnk4/zMRFTIuGxg8uhveAfeDK+36FQhRrmiPyBYBs4E7gE9TdcHNF/8qxh+uejshkprWOGY9j2PuqxFD/j4X4P+/28AqHWpEzRHtT3tDrbU3rpkkwMrdkfW3eZZM8BApJODH+fyj2L/91BDWSuG/vD+BMy6R6fWixqmwkCmlrtBa713r4dKKriNprV+OR2KJcseR0t1ApIii3bin3Yf1r18MYW134L1lJKHjTjYpMSHMVdkR2XXA3kJ2cwXbaCBlC9nEE3JokilHZCL5qS0bcU8ahiV/nSGus3PwDHqEcNsjTMpMCPNVWMi01meU+TrtlmsRIlVY/l6Ga/K9WAp3GeLhRs1LG/82bGpSZkIkh6juI1NKlduaSin1XXnxZJRXEOCFv1L+bgFRw1gXf4l73KCIIhZq35GS+56QIiYE0U/2qGjBovaxSiTehn9fYHYKQkRPa+wfz8Tx5gzUfquYB7qegu+mYeBwmpScEMml0kKmlHphz5eOMl/vdQjwRzySioe/dwcjYs0yZYqySELhEI7Xn8Qx/52IIf+5V+G/8AawRNtdToj0V9Un+YYKvtbAj8BbMc8oQbo2cHBKU/mLViQZnwfXjDHYlhh7DWiLBd+1Qwj2OsekxIRIXpUWMq31fVB6LUxr/UFiUkqMp0+qjd0i99uI5KF2bcc1eQTW1csNce3KwHvbaEIdjzcpMyGSW2X3kZ2otf56z8NCpVTP8rbTWi+MS2ZC1CBqw+rS6fXbNhvi4dr18A4ZT7hFW5MyEyL5VXZE9jz/TPJ4rYJtNNAiphkJUcNY//gZ17RRqBLjrNpQizZ4B49H16lvUmZCpIbK7iPrUObr5olJR4iaxfbVxzhfeAwVMk5GCnY8Hu/A0eCWxr9CVOWgpu0ppU4CQqnW/V6IpKE19vdexjn7xYihQK9z8V1zpzT+FSJK0d4Q/blSqseer+8C3gFmKaWGxTM5IdJSMIDzufHlFjHfpbfgu26IFDEhDkC0vy0dgb1dPG4BegGFwJfAhNinJUSaKi7E9cQD2Jb9ZAhrux3fzSMIdj3FpMSESF3RFjILEFZKtQZsWuulAEqpOnHLLIY2FodYXZiSS6iJNKK2bcI1cTjWjasNcZ1ZC8+gsYTbH2VOYkKkuGgL2TfAFKAJpQtssqeobY9TXjGjtabfx9vMTkPUcJZVf5Y2/i0wrlAebtAEz9BH0Y2amZSZEKkv2j431wFe4C/ggT2xw4HpccgpptYXh/irILI9VaZdboYWiWH9+RvcjwyKKGKhtkdQct8MKWJCVFNUR2Ra663APfvF3gfej0dSsfT9Fn9E7NSmThq4ZR0yEX/2T2fjeHU6SocN8cBxvfD9a4Q0/hUiBqKdtWhTSt2nlFqulCre8//7lFL2eCdYHX8XBLnpi50R8VdPqWtCNqJGCYdxvDED5ytTI4qYv+/l+AbcL0VMiBiJ9hrZBOBEYBCwBmgJjAJygaHxSa36PlzniYgdUduGyyanFUUc+by4nn0E22Jj9zatLPiuvpNgn/NNSkyI9BRtIbsU6KS13jtrYumexTaXkKSF7KetfkYt2h0Rv7JdpgnZiJpC7d6Ja8q9WP82rnCknS68A0cTOrqbSZkJkb6iLWRWILxfLEz0k0USrrzVoDvWsTPwiCwTshE1gcpfi3viMCxb8w3xcG7d0sa/LduZlJkQ6S3aQjQTmKOU6qOUaqeUOpXSafgz45da9fy+IxARkyIm4sXy5y9kjBkYUcRCzVrhuf8pKWJCxFG0R2R3Uzrt/nlK7yXbALwJPBinvKrlmWVFLNluLGTdGzq4rI3bpIxEOrN9+ynO5yaggsafueARx+K97UHIkD+ghIinaKff+4B79/yX9Kb9VhQRG3hEFkrJJA8RQ1pjf/81nDOfixgKnHQWvuuGgk16JgoRb5WeWtxzGnGhUmqHUupTpVRKrD2207//5Tw4um5S3ykgUk0wiPOFx8otYr4Lb8B34z1SxIRIkKqukT1B6WnE64BtlLapSjkv9a5Dsyz5UBEx4inGNWk49oXzDGFtteH9170Ezr8G5OhfiISp6tP9WKC51tqjlPoM+DMBOcXcqU3lxlMRG2r7FlyThmNdv9IQ1xlZeO4cS7jDMSZlJkTNVVUhc2itPQBa60KllMyWEDWWZU0erkkjsOwyNqEO12uEZ+gEdJOWJmUmRM1WVSFzKqXuL/PYvd9jtNYPxT4tIZKL9Zfvcc0YjfIau8WEWh+Gd9DD6JyUWNFIiLRUVSF7Gyh7A8zM/R7rmGckRJKxfTYH58tTUGHjJKJg5x54bx0FTpdJmQkhoIpCprW+OlGJCJF0wmEc//03jnlvRAz5T78Y/xX9wSKrKAhhNpnKJ0R5/D6cz47DvuhzQ1grhf/K2wicfpE5eQkhIqRdIVtTGKQkKGc8RTUU7sI9ZRTWFb8bwtrhxNv/PkKde5iUmBCiPGlVyHwhzVnztpqdhkhhatN63JOGYdm8wRAP59TGO2gc4dYdTMpMCFGRtCpkv+8IsLHEeEHebgG7RW5OFVWzLP8N99SRqCLj8j/hJi3xDBmPrt/YpMyEEJWJehkWpVRvpdQzSql39zzurJQ6+QCef6ZS6i+l1Aql1PBKtrtIKaWVUl2ife29/OHIU4r9WrlxWKWQicrlLluE+9EhEUUs2OEYSkY9IUVMiCQWVSFTSg2gtPP9OqD3nrAfeDjK51uBJ4GzgMOBK5RSh5ezXTZwJ/B9NK8bjSdOrB2rlxLpSGvsH7xBq3eeRQWM3esD3U/He/djkJltUnJCiGhEe0Q2FDhVaz2WfxbY/AM4LMrnHw+s0Fqv1Fr7KV0Cprz13scAEwBvlK9bqW4NHHI0JioWCuJ8aTLOt5+JGPJfcC2+f40AmzSbFiLZRVvIsoE1e77ee/7ORulRWTSaUno0t9f6PbF9lFKdKe3r+EGUrynEwfOU4JoyEvtncwxhbbXivXk4/n7XS+NfIVJEtJM9vgLuovRoaa+BwBexSEIpZQEmUdplPyp5eXkRsfUFFuCfLgser6fc7WqSmv79l8e+eyet35qObfM6QzzodLPq4v4UNWwDst8qJT9X0ZN9VbV27aq3gnq0hex24H2l1M1AtlJqKaVHY32jfP4GoHmZx832xPbKBo4EPt+z+GUjYI5S6jyt9eLyXrC8b3zbZh/89k9DV7fLTbt2KbGEWlzk5eVV+wck3VjW/o3r1cew7DDepuGvVYfgsIk0btbKpMxSh/xcRU/2VWJEu0L0hj2n/roDLSg9Tfit1joU5fssAtoppVpRWsAuB64s8/oFQL29j5VSnwN3VVTEhDgY1t8W4XriAZS3xBAPtWzPXxfcTCspYkKkpKjvI9Naa+DrPf8dEK11UCl1G/AxYAVe0FovVUo9BCzWWs+p/BWEqB7bFx/g/M/EyMa/x5yAt/99BNdtqOCZQohkF1UhU0qtooJO91rr1tG8htZ6HjBvv9j9FWzbK5rX3F8wXPU2oobRGses53HMfTViyN/nAvxX3S6Nf4VIcdEekd203+PGlF43i2wLbpItnhDnfrSt6g1FzRHw43z+UezffmoIa6XwXz6AwBkXy8xEIdJAtNfIFuwfU0otoPQIa0qskzoY76zyRMTkM6oGK9qNe+oorMt/NYS13YH31lGEuvQ0KTEhRKxVp9eiB4jqtGIibPdGnlc8pq7czFoTqS0bSxv/5hun14ezc/EOephw2yNMykwIEQ/RXiPb/1pWBnA28EnMM4qh+46tZXYKIsEsK5bimjISS+EuQzzcqHlp49+GTSt4phAiVUV7RLb/jRDFlPZO/E9Ms4mhEZ2yybBF3RNZpAHr4oW4nh6LChgbzoTaH4XnzrGQJX/YCJGOqixkexr+zgfe1lrHpAdiPBTLlMWaS2vsH8/E8eYMlDZOrg1064PvpmFgd5iUnBAi3qosZFrrkFJqutb65UQkdDC+2uRjxtJis9MQZgiHcLz2BI5PZ0cM+c+9Cv+FN4BFjsyFSGfRnlr8QCnVd8+9YEln+u9FETGZsFgD+Dy4ZozBtuQbQ1hbLPiuHUKw1zkmJSaESKRoC5kFeEcp9RWl7an2nb/RWt8Qj8QOxHZvZKesY+rKqaR0pnZtxzV5BNbVyw1x7crAe9uDhDoeZ1JmQohEi7aQ5QGPxTORWLqktZvTmjnNTkPEiWX9KlyThmPZvtkQD9epj3fweMIt2piUmRDCDJUWMqXUFVrrN7TW9yUqoVj412FZKLkbOi1Zl/2Ea/p9qBLjNdFQi7Z4B49D16lvUmZCCLNUdRU8culcIUxi++ojXI/fHVHEgkd1xXPvNCliQtRQVZ1alMMaYT6tcbz7HxzvvhQxFOh9Lr6r7wRrdZrUCCFSWVW//ValVG8qKWha6//FNiUhyggGcD7/GPZvIpvI+C69hUDfy6WpphA1XFWFzAk8T8WFTJNE/RZFmikuxDX9fmx//GwIa7sd3833Euza26TEhBDJpKpCVhztemNCxJLamo970nAsG9cY4jqrFp47HybcvqNJmQkhko1cWBBJx7LqT1yTR2Ap2GmIhxs2xTNkArpRM5MyE0IkI5nsIZKK9aevcT01BuU3tvUMtT0Sz6CxkJ1rUmZCiGRVaSHTWmcnKhEh7PPfwfHa9MjGv8f1wvevEeCQm9yFEJHk1KIwXziE442ncHwyM2LI3/cK/JfcLI1/hRAVkkImzOXz4np6LLafvjKEtbLgu+ZOgqecb1JiQohUIYVMmEYV7MA1ZSTWlX8Y4trpwjtwNKGju5mUmRAilUghE6ZQG9eUTq/fmm+Ih3Pr4R0yjnDL/RclF0KI8kkhEwln+XMJ7qmjUCXGdeRCzVrjHTIeXbeBSZkJIVKRFDKRULZv5uN8bgIqFDTEg0d0wXv7g+DONCkzIUSqkkImEkNr7HNewfnOCxFDgZ598V07BGzy4yiEOHDyySHiLxjE+dIk7AvnRQz5Lr6JwDn/J41/hRAHTQqZiK+SIlxPPIBt6Y+GsLbZ8d00jOAJp5qUmBAiXUghE3Gjtm/GNWk41vWrDHGdmY3njrGEOxxtUmZCiHQihUzEhWX18tLGv7u2G+Lh+o3xDJ2AbtzCpMyEEOlGCpmIOeuSb3HNeBDl26/xb5vD8A56BF2rtkmZCSHSkRQyEVO2Be/hfGUqSocN8eCxJ+G9ZSQ4XSZlJoRIV1LIRGyEwzjefgbHh29FDPnPvBT/ZbeAxWpCYkKIdCeFTFSf34fr2UewLfrCENbKgv//biNw2oUmJSaEqAmkkInq2b0L99SRWFcsNYS1w4V3wP2EOnU3KTEhRE0hhUwcNLVpHe6Jw7Bs2WiIh3Pq4B08jnCrQ03KTAhRk0ghEwfFsvxX3FNGoYp3G+KhJofgHToeXa+RSZkJIWoaKWTigNm+/x/OZ8ehggFDPHh4Z7y3PQiZ2SZlJoSoiaSQiehpjX3eGzjffjZiKNDjDHzX3wU2uwmJCSFqMilkIjqhIM6Xp2L/fG7EkK/f9QTOv0Ya/wohTCGFTFTNU4LrydHYfvvBENZWG74b7ibY4wyTEhNCCClkogpqxxZck0dgXfu3Ia4zMvHePobQ4Z1NykwIIUpJIRMVsqxdgWvScCw7txni4XoN8QyZgG56iDmJCSFEGVLIRLmsv/2A64kHUF6PIR5qdWhp49/cuiZlJoQQRlLIRATb5+/jfGkSKrxf499OJ+LtPwqcbpMyE0KISFLIxD/CYRyznsfx/msRQ/7TLsR/5UBp/CuESDpSyEQpvw/ncxOwf/8/Q1grhf+KAQTOuMSkxIQQonJSyAQUFeCeOgrr8t8MYe1w4r1lFKEuJ5mUmBBCVE0KWQ2nNm8obfy7eb0hHs7OxTv4EcJtDjcpMyGEiI4UshrMsmIp7in3ogoLDPFw4+al0+sbNDEpMyGEiJ4UshrKuugLXM88jAr4DfFQh6Px3D4GsmqZlJkQQhyYlC9kv+8IsHhroOoNRSmtsX/0No63nkZpbRgKnHAqvhvvAbvDpOSEEOLApXQh21gcosd7W8xOI3WEgjheewLHgncjhvznXY3/whuk8a8QIuWkbCHb5QvTdfbmcsfquy0JziYFeEtwzXgI2y/fGcLaasV33VCCPfualJgQQlRPyhay77f4KQzoiPgVbTNomSU37Zaldm4rbfy7Js8Q1+5MvLc9SOjILiZlJoQQ1ZeyhcwTjCxipzV18tRJtU3IJnlZ1q/ENXE4lh3GU7DhOg3wDhlPuHlrkzITQojYSNg5OKXUmUqpv5RSK5RSw8sZH6KUWqaU+lUptUAp1bKy17vu8x0Rsdf6SCPbsrJXLsM99vaIIhZq2Q7P/TOkiAkh0kJCCplSygo8CZwFHA5coZTa/07bn4EuWuujgJnAowfyHue1dOGwykSFvWwLP6TNm9NQnmJDPHh0Nzz3TkXXrmdSZkIIEVuJOrV4PLBCa70SQCn1JnA+sGzvBlrrz8ps/x1w1YG8weG17TFIMw1ojWP2izjeezliKND7PHxX3wHWlD2jLIQQERL1idYUWFfm8XqgayXb3wh8eCBvcEfHrINIK80E/DhfeAz7N/MjhnyX3UrgrMtker0QIu0k3Z/mSqmrgC7AydE+5/mjvGxY9Xf8kkoBVk8xrWY+hX3NX4Z42Gpjzfk3sqvdsbBihUnZpYa8vLyqNxKA7KsDIfuqau3atavW8xNVyDYAzcs8brYnZqCUOhUYCZystfZF88K3HJbJRcc2jUmSqUptzcf9/Bgs+WsN8UBGFoEh46nf7kjqm5RbqsjLy6v2L1NNIfsqerKvEiNRsxYXAe2UUq2UUg7gcmBO2Q2UUp2AZ4DztNZRt+sYcETNPqVo+fsP3A8NiChi4YbNWH7dCMLtjjQpMyGESIyEHJFprYNKqduAjwEr8ILWeqlS6iFgsdZ6DvAYkAX8V5Vex1mrtT4vEfmlKuuPX+J6eizKbzx4DbU7Es+gh/HnS/suIUT6S9g1Mq31PGDefrH7y3x9aqJySQf2T2bieP3JyMa/XXvju2k4OJyAFDIhRPpLuskeogrhEI7XZ+CYPytiyH/2lfgvvgks0mtSCFFzSCFLJT4PrqfHYvvpa0NYWyz4rhlMsPe5JiUmhBDmkUKWIlTBDlyT78W66k9DXLvceAeOJnRUZbflCSFE+pJClgLUxjW4Jw7Dsm2TIR6uXa+08W+LtiZlJoQQ5pNCluSsf/yMa9p9qJIiQzzUvA3eIePQdRqYlJkQQiQHKWRJzPb1JziffxQVChriwY7H4R04GtyZ5iQmhBBJRApZMtIa+3sv45z9YsRQ4OSz8V0zGGzyTyeEECCFLPkEAzhfnIj9q48ihnyX3Ezg7Cul8a8QQpQhhSyZFBfieuIBbMt+MoS1zY7v5uEEu/UxKTEhhEheUsiShNq2Cdek4Vg3rDbEdWYtPHeOJXzoUeYkJoQQSU4KWRKwrPoL1+QRWAp2GOLh+k3wDB2PbtzCpMyEECL5SSEzmXXJN7iefAjl9xrioTaH4xn0CNTKNSkzIYRIDVLITGRb8C7OV6ahdNgQD3bpifeWkXsa/wohhKiMFDIzhMM43noax0dvRwz5z7oM/6W3SONfIYSIkhSyRPP7cD3zMLbFCw1hrSz4rr6DYJ8LTEpMCCFSkxSyBFK7d+KaMhLr38sMce1w4R14P6FjupuUmRBCpC4pZAmi8tfinjgcy9aNhng4pw7eweMItzrUpMyEECK1SSFLAMtfv+KeOhJVXGiIh5oegnfoBHTdhiZlJoQQqU8KWZzZvl2A87nxqGDAEA8e3hnv7Q9BRpZJmQkhRHqQQhYvWmN//3WcM/8dMRTocSa+64eCzW5CYkIIkV6kkMVDMIjz5cnYv/ggYsh34Q0EzrtaGv8KIUSMSCGLNU8xridGY/t9kSGsrTZ8N95D8MTTTUpMCCHSkxSyGFI7tuCaOBzr+pWGuM7IwnvHGEKHdTIpM5GqtNYUFRURDoer3jhBXC4XBQUFZqeREmRfRbJYLGRlZaFieFZKClmMWNbk4Zo0AsuubYZ4uF4jPEMnoJu0NCkzkcqKiopwOp04HA6zU9nH6XTicrnMTiMlyL6K5Pf7KSoqIjs7O2avKYUsBqy/fI9rxmiU12OIh1p1wDv4EXROHZMyE6kuHA4nVRETorocDgcej6fqDQ+AFLJqsn02F+fLk1H7nfoJdj4R7633gVP+GhNCiHiSzrQHKxzG8fYzuP4zMaKI+U+/qPQeMSliIk28//775Obmsnz58n2xL7/8kssuu8ywXf/+/XnvvfcACAQCjB49ms6dO9OzZ09OO+005s+fX+1cJk2aRKdOnejSpQsLFiwod5svvviCnj170qNHD84880xWriy9bj1ixAh69OhBjx49OPbYY2nR4p+1/l5//XU6d+5M586def311wEoLCzct32PHj1o3bo1w4cPB2Dt2rWcd955dO/enbPPPpsNGzbse61169bRr18/TjrpJLp27cqaNWsAePbZZ+nUqRO5ubls37593/Zaa+655x46depE9+7dWbJkCQALFy40vH/Dhg15//33AVi9ejV9+vShU6dOXH/99fj9/kq/x19//ZXTTjuNbt260b17d955550q99de7733Hrm5ufz888+G+Lp162jatCnTp0/fF3vqqac44YQT6NatGzNmzKj8HzNGUvqITAE5DhNqsd+H89/jsf/wmSGslcJ/5UACp1+c+JyEiKNZs2ZxwgknMHPmTIYMGRLVcx5++GE2bdrEt99+i9PpZMuWLXz99dfVyuPPP/9k1qxZfPfdd+Tn53PBBRfw448/YrVaDdsNGTKE119/nUMPPZTnnnuOxx57jKeeeopx48bt2+aZZ57h119/BWDnzp1MmDCBzz//HKUUJ598Mn379iU3N5evvvpq33NOPvlkzj33XADuu+8+Lr/8cq688kq++OILHnzwQZ599lkAbr31Vu666y5OOOEEgsEglj2rWXTt2pUzzjiDc845x5Dv/PnzWblyJT/99BOLFy9m6NChLFiwgJ49e+57/507d9KpUydOOeUUAEaPHs2AAQO46KKLGDx4MK+88go33nhjhd9jRkYGTz/9NG3atCE/P59evXpxyimnkJubW+H+gtJi/vTTT9OlS5eIf4+RI0dy6qmn7nu8bNkyXn75ZRYsWIDD4eCiiy7izDPPpHXr1lH/Gx+MlD4iu7FDJrnOBH8LhbtwPzo0sog5nHjvGCNFTKSdoqIivvvuO6ZPn274K74yJSUlvPTSSzz66KM4naXr6jVo0IB+/fpVK5d58+Zx0UUX4XQ6OeSQQ2jdujU//vhjxHZKKQoLS1vC7d69m8aNG0dsM2vWLC6+uPT3dcGCBfTu3ZvatWuTm5tL7969+fTTTw3br1ixgm3bttG9e2lz77/++ouePXPHRIQAABa5SURBVHsC0LNnTz788EOgtNiGQiF69+4NQFZWFhkZGQAcffTRtGwZOfFr3rx5XH755SilOO644ygoKGDTpk2Gbd577z1OO+00MjIy0FqzcOFCzj//fACuuOIKPvgg8r7Vst9j27ZtadOmDQCNGzemXr16+44KK9tfDz/8MIMGDdr377jX+++/T8uWLenQocO+2PLlyzn22GPJyMjAZrNx4oknMnfu3Ii8Yi2lj8ge65aT0PdTm9eXNv7dvN4QD9eqXdr4t3WHCp4pRGzkvrih6o0OwK7rm1a5zbx58+jTpw9t27alTp06/PLLL3Tt2rXS56xcuZJmzZpRq1atKl9/xIgRfPnllxHxvUcaZeXn5xuODJo0aUJ+fn7Ec6dNm8Yll1yC2+0mOzs74pTm2rVrWbNmzb5ClJ+fT9Om/+yL8l531qxZ9OvXb9+08SOPPJK5c+fSv39/5s6dS2FhITt27GDFihXk5ORw1VVXsXr1anr37s3o0aMjjhr3/77Ke/9GjRoZ3n/gwIEA7Nixg5ycHGw2W4X57v89lvXjjz8SCARo1apVpftryZIlbNiwgTPOOINp06bte35RURFTp05l9uzZhtOKhx12GGPGjGHHjh24XC7mz5/PMcccU+H3HSspW8ja1rLF9D6Eqljyfsc95V5U0W5DPNykJZ4h49H1I//iEyIdzJo1i1tvvRWACy+8kNmzZ9O1a9eY/f6VPRUWKzNmzOC///0vXbp0Ydq0aYwcOdLwgfvOO+9w3nnnVVpc9vfOO+/wzDPP7Hs8ZswY7r77bl5//XW6d+9OkyZNsFgshEIhvv32WxYuXEi9evXo378/r732Gtdcc81Bfz+bNm1i2bJl9OnT54DyLe973LRpE7fccgtPPfXUvlOe5e2vqVOnMnLkyHKvc40fP54BAwaQlWXsFXvooYdy55130q9fPzIyMujYseMB7eODlbKFzJ7AM4rWHz7H9ezDqMB+jX87HIP3jjGQGbv7IYRIJjt37mThwoUsXboUpdS+G7PHjRtHnTp12LVrV8T2devWpXXr1qxfv57du3dXeVR2IEdkjRs3Nkyq2LhxY8Rpw23btvH777/vO3Lr16/fvtNre82aNYvHH3/c8Lplr4Vt3LiRHj167Hv822+/EQwGDUcXjRs35tVXXwVKj1Dmzp1Lbm4uTZo04cgjj+SQQw7B6/Vy9tlns3jx4kr3QVXf1+zZsznnnHOw20v7s9apU4eCggKCwSA2m63c/bD/9wilpw0vvfRS7vv/9u49uqrqTuD495eQBzFAUKSKEp1RccACEwgEYQkoJUBUlLGWN1hALGJX8QFSFAYoAgNWBBYMo+hoK0IrWQRwVOoYK6UIE5SliVIEEVq0JTxMIMBNQvKbP87hem8eNxcJ95H8PmtlrWSfc8/+nZ2b+8s+Z5+9Z86kW7duAdvr1KlT7Nmzx3s/r7CwkOHDh7N27Vo++ugjNm7cyKxZsyguLiYmJoaEhAQmTpzImDFjvEl77ty5tGnTJuC514eovUfW75oQjAhUJe6tdTRdMbtaEivv2R/PE4ssiZkGbePGjQwdOpSCggLy8/P57LPPaNu2Ldu3b/cOGti7dy/gXMoqKCigY8eOJCUlMXr0aKZPn+4dTXfs2DFycnKq1bFgwQK2bdtW7atqEgMYNGgQ2dnZlJaWcvDgQb788ku6du3qt09KSgonT55k//79ALz//vu0a9fOu/2LL76gqKiI7t27e8v69etHbm4uRUVFFBUVkZub69f7yc7O5r777vOr5/jx497EvmTJEkaOHAlAly5dKC4u5tgxZ3KErVu3cvPNgdcbHDRoEOvWrUNVycvLo3nz5tUuK/rWLyLcdttt3hGia9euJSsrK+A5lpWVMWrUKIYNG+a9txaovVq0aMGBAwfIz88nPz+f9PR01q5dS1paGm+//ba3fNKkSTz++ONMnDgRgKNHjwLOiMbNmzdX+yfiUojaHtns9LqvvV+UinMk/HYZce9vqrap7J6xlA15wCb+NSEXzD2t+rR+/XqmTJniV3bXXXeRnZ1Nr169eOGFF5g8eTIej4e4uDiWL19OixbOveunn36aefPmkZGRQWJiIklJScyYMeOi4mnfvj1DhgwhIyODJk2a8Oyzz3ovXd1///0sW7aMq6++mqVLlzJmzBhEhJSUFFasWOE9xvmk4HtptGXLlkydOtU7QGPatGm0bNnSu33Dhg288cYbfrFs27aNOXPmICL07NnT2/uJjY1l3rx5DB48mMrKStLS0hg7diwAq1atYtmyZRw5coRevXrRv39/li9fTmZmJu+++y5paWkkJSX5xXvo0CG+/vprvx4iwJw5cxg3bhzz5s2jU6dOjB49OuA5btiwge3bt3PixAnv4wUrV66kU6dOAdvrQo0ZM4YTJ054fz8pKSnf+1jBElW95JXUl+Li4tAE6zlD4sq5NPlkh1+xxsZS+tMnOHfboJCEcbH27dvHTTfdFO4wokKktlVxcbE3MUQKj8dj0y4FydqqZoHe1y1atLjgHkLU9sguFfn2GIlLfknsoX1+5dr0Mjw/n0vFLV1reaUxxphwsETmI+ZvB0h87kliThz1K6+84gd4HltA5bWX9qE+Y4wxF84SmSu2YBeJy2chnjN+5RXXtcPz2AI05YowRWaMMSYQS2RAk61vkfDKr5GKCr/yc5174Hl4FiQmhSkyY4wxdWnciUyV+OyXiN/8WrVNZf3upWzkIxDbuJvIGGMiXeP9lC4vI+GlRcR96D+fmopQNmwS5QPut+H1xhgTBaL2geiLUnKSpounVk9icfF4Js+mfOBPLIkZ46OwsJAJEybQuXNnMjMz6d+//yWfDHb37t1MmzbtktZhGoZG1yOTwm9o+tyTxPz9b37l2qwFZ6fMp/LGW8IUmTGRSVUZOXIkw4cPZ/Xq1Xg8HgoLC72zvV8qaWlppKWlXdI6TMPQqHpkMV9+TtO5D1dLYpVXteXMzJWWxIypwdatW4mLi2PcuHHestTUVB566CHWrFnD1KlTveVDhw71zpuYm5tL//796d27N2PHjqWkpARw1tHKyMigZ8+ePP300wDk5ORw66230qtXLwYNciYc8F24c8GCBUyePJk777yTzp07s2rVKm+dixYtIj09nYEDBzJ+/Hi/yYFN49BoemSxu/5E4qpfIeVlfuUV7Tpy9hfzIDmyZk8wpibJY/vW6/FKXv1jnfvs2bOHzp07X9Bxjx8/zuLFi8nJyeGyyy7j+eefZ8WKFTz44IO8+eab5OXlISLeSYcXLVpEdnY2bdq0qTYR8Xn79u1j8+bNlJSUkJ6ezvjx48nPz2fTpk1s27aN8vJy+vTpE5JlQ0xkafiJTJW4LeuJX7cSqTIdV3nGHZROeBLiE2p5sTGmqunTp5OXl0d8fDwTJkyocZ+8vDz27t3LgAEDACgvL6dbt240b96chIQEHnnkEQYMGMDAgQMBZ+Xkhx9+mCFDhnhXYK4qMzOThIQEEhISuPLKKyksLGTHjh1kZWWRmJhIYmKi93imcWnYiayygvjXVxD/bvVVbcvuHkXZv42DmEZ1ddWYC9a+fXs2bfpu8uyFCxdy+vRp+vbtS5MmTbwzwIMztyA499Vuv/12XnrppWrHy83N5YMPPmDjxo28+OKLbN68mSVLlrBr1y62bNlCnz59+OCDD6q9zneF4tjYWM6dO1efp2miWMP9FC89S+LSmdWSmMbE4PnpE5T9eIIlMWOC0Lt3b0pLS/2S0pkzzgw4qamp5OfnU1lZyeHDh/n4448B6NatGzt37uTAgQMAnD59mv3791NSUsLJkyfJzMxk/vz5FBQUAPDVV1+Rnp7OU089RatWrTh8+DDB6NGjB++88w4ej4eSkhK2bNlSn6duokSD7JFJ0XFn4t+DX/iVa2ISnkdmU9Gxey2vNCayBXNPq76JCGvWrGHGjBksXbqUyy+/nOTkZObMmUOPHj247rrryMjIoF27dnTq1AmAVq1asWLFCsaPH09paSngLOuSnJzMiBEjvD23Z555BoCZM2dy4MABVJXevXvTsWNHv4Uua9OlSxcGDRpEr169aN26NR06dKhzIU/T8DS4ZVxiDn9F4pLpxBw74lde2bIVnscWUpl64yWLL9JE6tIkkShS28qWcalbSUkJycnJnDlzhqysLJ5//vmIGfARaW0VKWwZlwBiP/+YxOUzkTOn/corUm/A8+hC9PIrwxSZMeZSmTJlCn/5y18oLS1l+PDhEZPETOg0mETWZNsWEl5ejFT43wA+17E7nsmzoalN/GtMQ7R69epwh2DCLPoTmSpxOa+SkPNKtU3lfe+mdMwvbOJfY4xpwKL7E/5cOQkvP0vcn6uPVCr9yUOUZw2zORONMaaBi95EdvoUictn0WTPbr9ijYuj9MFfci7jjjAFZkz9iYmJoaysjPj4+HCHYky9KCsrI6aeH32K2kTWdN7Pif3moF+ZXtacs1PmUdmuU3iCMqaeJScnU1JSwtmzZ8MditfJkydtiHuQrK2qi4mJITk5uV6PGbJEJiIDgaVALLBaVRdW2Z4A/AboChwHhqrqwdqOVzWJVbZuw9nHF6FXXVu/gRsTRiJCs2bNwh2Gn8LCQtq2bRvuMKKCtVVohGRqCxGJBVYAg4AOwHAR6VBlt/HAt6p6I7AE+I9gj19x4y2cmbnSkpgxxjRCoZqjqTuwX1UPqGoZsA64p8o+9wCvut+vB/qJ1D1So7xbX84++Rw0T6nXgI0xxkSHUF1avAbwXQTsMJBR2z6qek5EioErgGM1HTAcU/VEm0icqSJSWVsFz9oqeNZWoWGz5hpjjIlqoUpkXwO+dzyvdctq3EdEmgAtcAZ9GGOMMbUK1aXFPOAmEfknnIQ1DBhRZZ9NwFjgQ+DHQK5WmdH4+0wmaYwxpmELSSJz73k9AmzBGX7/sqp+JiJzgV2qugl4CfitiOwHTuAkO2OMMSagkN0jU9W3VLWdqt6gqs+4ZbPcJIaqelT1fnf4/SzgbRHZLyLTqx5LRBJE5Hfu9p0icn2oziPSiMhAEdkboK0eE5HPReRTEXlPRK4LR5yRoq728tnvPhFREUkPZXyRJJi2EpGfuO+vz0Tk9VDHGCmC+DtMFZH3RWS3+7eYFY44I4GIvCwihSJSUMt2EZFlblt+KiJd6jyoqkbUF06P7Uvgn4F44BOgQ5V9HgZWud8PA34X7rgjuK1uB5Lc7yc11rYKtr3c/ZoBW4EdQHq4447UtgJuAnYDLd2fW4c77ghuqxeASe73HYCD4Y47jO3VG+gCFNSyPQt4GxCgB7CzrmNG4qjFS/bMWQNUZ1up6vuqesb9cQfOQJvGKpj3FsCvcB7I94QyuAgTTFs9CKxQ1W8BVLUwxDFGimDaSoHzc1W1AL4JYXwRRVW34tw+qs09wG/UsQNIEZGrAx0zEhNZTc+cXVPbPqp6Djj/zFljE0xb+RqP859OY1Vne7mXMdqq6v+EMrAIFMx7qx3QTkT+LCI73GnoGqNg2mo2MEpEDgNvAT8PTWhR6UI/16J30mBzYURkFJAO9Al3LJFKRGKA54AHwhxKtGiCc3mxL05Pf6uIdFTVorBGFZmGA6+o6q9F5FacgW0/VNXKcAfWEERij8yeOQteMG2FiPwIeAoYrKqlIYotEtXVXs2AHwJ/FJGDONfnNzXSAR/BvLcOA5tUtVxVvwK+wElsjU0wbTUe+D2Aqn4IJAKtQhJd9Anqc81XJCYy7zNnIhKPM5hjU5V9zj9zBrU8c9ZI1NlWIpIG/BdOEmus9zDOC9heqlqsqq1U9XpVvR7nnuJgVd0VnnDDKpi/wxyc3hgi0grnUuOBUAYZIYJpq78C/QBEpD1OIjsa0iijxyZgjDt6sQdQrKp/D/SCiLu0qPbMWdCCbKvFQDLwhjse5q+qOjhsQYdRkO1lCLqttgCZIvI5UAFMVdVGd2UkyLZ6HHhRRB7FGfjxQCP95xsRWYvzD1Ar957hvwNxAKq6CuceYhawHzgD/LTOYzbStjTGGNNAROKlRWOMMSZolsiMMcZENUtkxhhjopolMmOMMVHNEpkxxpioZonMNEoi8pqIzA53HHVxZ1S/LcD2P4jIyFDGZEyksURmopqIHBSRsyJS4vPVJkyxvCYiZW4MJ9wk0+5ijqmqN6vqn9zjzxORV6psz1TVNRdTR1Ui0sRdwua0ey6HRWSxO4VXMK//kTszijEhYYnMNAR3q2qyz1c4Zxafr6rJOFPsnABeDmMsF+sW91zuAEbz3Ww6xkQUS2SmQRKRGBFZLyL/EJEiEfmjOzVQTfu2FpG33P1OiMhWn23XisgGETkqIl+JyORg6lfV08BanLkbEZFEd7HAv4vI1yLynDudUV31HxaRviJyFzANGOn2kj5yt28TkQdEpKmInBSRf/F57VVub/UK9+fBIvKJW882EflhkOfyBbAd+FefY08QkT0ickpEvhSRCW55C2AzkOrTQ27t/j5muPseE5F1ItIymPqNqYslMtOQvYkzie1VQAHw21r2m4ozR+CV7r5Pg3c2/Ddx5tK7BugPTBWRfnVVLCLNgBE4C0+Cs+p5OtAJSAN6Ab8MVL8vVX0TWASscXudXatsP4sz9+Fwn+KhwHuqelxEugEvAhNwljx6Gdh4PpnWcS7t3Xj3+xQfAe7EWWPrQWC5iHRS1WLgbpyp0M73kAuBR939e+NMAlsCLKurbmOCYYnMNAQ5bi+jSERyAFS1UlVfUdVTqurBWQ+qq4hcVsPry4E2QKqqlrkL/wHcCjRX1flu+X6ceT4Dze05XUSKcGaCTwDGueUjgdmqetT9YJ+Lc7kuUP0X6nX8E9kItwxgIrBSVfNUtUJVz1/y7BbgeJ+KyGngc+BdnMmnAVDVze5CkqqqucB7QK2DUoCfATNU9Wv39zEHuD/Y+27GBGJvItMQ3KuqKe7XvQAiEisii0TkgIic5LveRE1LZywEDgHvuZe+prrl1+FcIjufJItwLu9dFSCWhW4cV6vqve7yJuAkqkM++x3iu8UCa6v/Qv0vzmq6XUXkBqADsNHnXJ6sci5XE3jBwk44S9uMwEnq3n8CROQuEdnpXgotAjIJvCxJKrDZp+58t7z1hZ+mMf4skZmGagzODNp34KxXd6NbLlV3VNWTqvqou3TLvTgf+H1wVqnd55MkU1S1mare/T3i+QYnmZyXirvGUoD6q4UaqAJ3tfQ3cHplI3DWCjvtbv4bMKfKuSSp6u/rOGalqq4FduGsaYeINAXWAwuAH6hqCvAHvmvbmuI8DPSvUn+iqv4jUP3GBMMSmWmomgGlOAuuJgHP1LajiNwtIjeIiADFOEuSVAIfAmUi8rg7WCNWRDqKSNfajhXAWmCWiLQSkSuBmcBrddRf1RHgene/2ryOc2/M97IiOPfHJotIN3Eku/XWdKm1JguBn7mxJwDxOOtpVbgDUXzvGx7BWaKjmU/ZKmC+iKS659xaRBrlckKm/lkiMw3Vf+P0gr4BPsMZdVebm4FcnAEIfwaWquqf3B5OFtAdOAgcw7lP1Px7xDMH+ARn0MmnwE6cHk2t9ddwjN/hJJATIvJ/tdSzHTiHM3DkD+cLVXUHMAn4T+BbnHt4o4INXlV34yT2J1S1CGfwxgacRwx+jDMo5vy+BUA2cNC9lNgaeA54B+fy6Sk3zkD354wJmq1HZowxJqpZj8wYY0xUs0RmjDEmqlkiM8YYE9UskRljjIlqlsiMMcZENUtkxhhjopolMmOMMVHNEpkxxpioZonMGGNMVPt/rwjRQSwXzg4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5693a33320>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "# Train Neural Network\n",
    "sk_net.fit(x_train, y_train)\n",
    "\n",
    "print('#### Neural Network ####')\n",
    "print(evaluate(sk_net, x_train, y_train))\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(sk_net.predict_proba(x_train)[:, 1], y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#### Ensemble #####\n",
      "F1       = 0.9050632911392406\n",
      "ROC AUC  = 0.7982565111110425\n",
      "Matthews = 0.6328586894481488\n",
      "Accuracy = 0.8584905660377359\n",
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687c23cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 472 ms, sys: 280 ms, total: 752 ms\n",
      "Wall time: 424 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Train Soft Probability Model\n",
    "soft_vote = PreTrainedVoteEnsemble(trained_estimators=[('nn', sk_net), ('rf', forest)])\n",
    "\n",
    "print('#### Ensemble #####')\n",
    "print(evaluate(soft_vote, x_train, y_train))\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(soft_vote.predict_proba(x_train)[:, 1], y_train)\n",
    "graph.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#### Forest (Testing) ####\n",
      "F1       = 0.8858131487889274\n",
      "ROC AUC  = 0.7740740740740741\n",
      "Matthews = 0.6100736078313957\n",
      "Accuracy = 0.835\n",
      "None\n",
      "\n",
      "{'max_depth': 20, 'max_features': 15, 'min_samples_split': 25, 'n_estimators': 400}\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687c1d400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#### Neural Network (Testing) ####\n",
      "F1       = 0.8860759493670887\n",
      "ROC AUC  = 0.7478882391163093\n",
      "Matthews = 0.46216382345751306\n",
      "Accuracy = 0.82\n",
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687bea6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#### Ensemble (Testing) ####\n",
      "F1       = 0.898360655737705\n",
      "ROC AUC  = 0.7801729963508582\n",
      "Matthews = 0.572668780490686\n",
      "Accuracy = 0.845\n",
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f5687b5c710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Forest\n",
    "print('#### Forest (Testing) ####')\n",
    "print(evaluate(forest, x_test, y_test))\n",
    "print()\n",
    "print(forest.named_steps['gridsearchcv'].best_params_)\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(forest.predict_proba(x_test)[:, 1], y_test)\n",
    "graph.show()\n",
    "\n",
    "# ANN\n",
    "print('#### Neural Network (Testing) ####')\n",
    "print(evaluate(sk_net, x_test, y_test))\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(sk_net.predict_proba(x_test)[:, 1], y_test)\n",
    "graph.show()\n",
    "\n",
    "# Ensemble\n",
    "print('#### Ensemble (Testing) ####')\n",
    "print(evaluate(soft_vote, x_test, y_test))\n",
    "\n",
    "graph.figure(figsize=(6, 6))\n",
    "plot_roc_curve(soft_vote.predict_proba(x_test)[:, 1], y_test)\n",
    "graph.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# NOW! Asking it Questions\n",
    "\n",
    "Obviously what we've all been waiting for"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "<b>Education</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(\"bachelor's degree or higher\", 0), ('some college', 3), ('less than high school', 2), ('high school', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Marital Status</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('living with partner', 1), ('never married', 3), ('separated', 4), ('married', 2), ('divorced', 0), ('widowed', 5)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Partner Education</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('some college', 3), (\"bachelor's degree or higher\", 0), ('high school', 1), ('less than high school', 2)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Gender</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('female', 0), ('male', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>House</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('a building with 2 or more apartments', 0), ('a one-family house detached from any other house', 3), ('a one-family house attached to one or more houses', 2), ('a mobile home', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Income</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('$20,000 to $24,999', 7), ('$40,000 to $49,999', 11), ('$125,000 to $149,999', 3), ('$15,000 to $19,999', 4), ('$175,000 or more', 6), ('$75,000 to $84,999', 16), ('less than $5,000', 18), ('$100,000 to $124,999', 1), ('$60,000 to $74,999', 14), ('$85,000 to $99,999', 17), ('$25,000 to $29,999', 8), ('$50,000 to $59,999', 13), ('$35,000 to $39,999', 10), ('$7,500 to $9,999', 15), ('$150,000 to $174,999', 5), ('$12,500 to $14,999', 2), ('$30,000 to $34,999', 9), ('$5,000 to $7,499', 12), ('$10,000 to $12,499', 0)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Msa</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('metro', 0), ('non-metro', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Rent</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('rented for cash', 2), ('owned or being bought by you or someone in your household', 1), ('occupied without payment of cash rent', 0)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Political</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('democrat', 0), ('republican', 2), ('other', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Religion</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('catholic', 2), ('jewish', 5), ('other non-christian, please specify:', 10), ('protestant (e.g., methodist, lutheran, presbyterian, episcopal)', 12), ('other christian', 9), ('none', 8), ('pentecostal', 11), ('baptist-any denomination', 0), ('eastern orthodox', 3), ('mormon', 6), ('muslim', 7), ('hindu', 4), ('buddhist', 1)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Work</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('working - as a paid employee', 5), ('working - self-employed', 6), ('not working - looking for work', 1), ('not working - retired', 4), ('not working - disabled', 0), ('not working - other', 3), ('not working - on temporary layoff from a job', 2)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Gender Older</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(1, 1), (0, 0)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "<b>Education Difference</b>"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[(1, 1), (0, 0), (2, 2), (3, 3)]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# TEMP input key\n",
    "for text_col in split_for(feature_names, False):\n",
    "    display(Markdown(f\"<b>{text_col.replace('_', ' ').title()}</b>\"))\n",
    "    print(list(zip(df[text_col].unique(), label_encoders[text_col].transform(df[text_col].unique()))))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['age', 'partner_age', 'age_diff_abs', 'children', 'visits_relatives', 'education', 'marital_status', 'partner_education', 'gender', 'house', 'income', 'msa', 'rent', 'political', 'religion', 'work', 'gender_older', 'education_difference'] 18\n"
     ]
    }
   ],
   "source": [
    "print(feature_names, len(feature_names))\n",
    "def ask(\n",
    "    age, partner_age, age_diff_abs, n_children, visits_relatives, education, marital_status, \n",
    "    partner_education, gender, house, income, msa, rent, political, religion, work, gender_older, education_difference):\n",
    "    # Initlise\n",
    "    vector = [age, partner_age, age_diff_abs, n_children, visits_relatives, education, marital_status, \n",
    "        partner_education, gender, house, income, msa, rent, political, religion, work, gender_older, education_difference]\n",
    "    \n",
    "    vector = np.array(vector).reshape(1, len(vector))\n",
    "    junk = np.vstack([vector, vector, vector])\n",
    "    \n",
    "    # Explanation\n",
    "    explainer = LimeTabularExplainer(\n",
    "        x_train, \n",
    "        class_names=['Broken Up', 'Stayed Together'], \n",
    "        feature_names=list(feature_names), \n",
    "        categorical_features=split_for(feature_indices, False), \n",
    "        categorical_names=split_for(feature_names, False), \n",
    "        discretize_continuous=True\n",
    "    )\n",
    "    for m, n in zip([forest, sk_net, soft_vote], ['Forest', 'Neural Network', 'Ensemble']):\n",
    "        explanation = explainer.explain_instance(junk[0, :], m.predict_proba, num_features=15)\n",
    "        font = '#' if n == 'Ensemble' else '###' \n",
    "        display(Markdown(f'{font} {n} Probability of Relationship Success = {explanation.predict_proba[1] * 100:0.1f}%'))\n",
    "        explanation.as_pyplot_figure()\n",
    "        graph.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "### Forest Probability of Relationship Success = 72.3%"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56a2929cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "### Neural Network Probability of Relationship Success = 94.8%"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgEAAAEGCAYAAAD8C0ZEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzs3Xm8XdP9//HXW4hZYizl6xs1x1DzPESj5qJFo4aKUnN1+FL6rSqqmpZftabWUBI1li9tiiI1FDFFiAyIhEQpNSeKouTz+2OtIzsn59x77nTOPTnv5+NxH/ecvdde+7PPOffutddeZ30UEZiZmVnrma/RAZiZmVljuBFgZmbWotwIMDMza1FuBJiZmbUoNwLMzMxalBsBZmZmLcqNALNM0lBJHzc6DgBJgySFpJUaHUuJpOmSTqnTvgZJmijpP5Lurcc+y/Yfkg6q936bXW/83HaGpGGSJjY6jnpwI8B6BUnDJf210XEYSLqsyol3U+DcOoXxG+Bx4HPAV+q0z15F0p6SHpD0lqT3JE2VdLWkJfL6bfIJd0BjI61djretn+l1jme1vN8t6rnf3sSNADOrSUS8HhHv1Wl3qwOjIuLFiHirMxVImk9Sn26Oqy4kfQG4CRgFbAOsDxwLvAMs2MDQumqFws8+edlGhWWbNiiuupDUR1KvOu/2qmDMqpG0uKSLJb0u6UNJj0naqazMcpKukPSqpA8kTZb0jbxOki6V9Jykf0t6XtJZkjr8D1XStyQ9k/cxRdIPJc2f160m6R1J3y2UXztfyR2Rnw+V9LGkHSVNyvU8ImmDNvbZbvySTstXi3vl+N6TdK+k1QtllpR0laS/53omS/ofSSrVARwGbF+4Ohua181xO6C990TSgLz9VyXdIun9HPfQNo5zkKQA+gBXlu1/C0n35bjflnSNpOUqHP8QSc8AHwFrVNnPYpJ+JenFHPt0Sf/bRlzfljRO0ruS/inpOkkrFNYvIOmXkl7K9b0i6brC+nUk3SFpRn5fnpZ0cLX9AXsC4yLi9Ih4KiKei4g7IuLoiHhd6er//lx2Wn6d7s372kjSXyS9luMdI2mXstdpcoVjvFzSXYXnG0u6M9fxuqSbJP132Tbfysf8vqQ7gJXbOCYi4p+lH6DUuHu9sPz1XG9/Sb+T9Ebh72OHsn1vmo/tw/w53iu/NycUyiwh6cL8fryXP6NfyusWAqbkog/l1/CZsn3sJ+nZ/BrcJWmVsvWb5+XvKf3f+YMKt0KUbytIOkjSs8CHwByvYcNFhH/80/AfYDjw1zbW3wBMB3YG1gZ+Tfonv1ZevzDwNKkLeUdSN/JOwP55/XzAT4HNgQGkf7KvAKcX9jEU+LidOE8DXgC+DKwC7Ab8HfhJocyBpD/2jYCFgPHAH8r2MyvHuj3pKu8W4B/AwrnMICCAlToQ/2nAe8DtwMbA54GxwP2FMssDJ+fYVgEOAt4FDs3rFwOuBh7MZZcvxDQdOKUD78mAfAzPA18FVgPOAj4G1qjy+vbN+wzSle/y+b1dnnQVfA2wHunqeDxwX9nxvw/8Lb9OawCLV9iHgHtzXHuTPivbAd8slAngoMLzb5M+V6sAW+bX52+F9d8DXsrv28qkK9rvFNaPz7EPzPvbFdijjc/ZScAMYLMq6/vkz0DkfS0PLFX47AwF1smvwZn5fVkjr18pvwfbF+pbPH8OhuTnA/Pz04G18mt+A/AssFAus1eu53t5P4cBr1L43LbztzSoWlngz8Bz+TUfSLo99AHwubx+CeB1Um/JesBWwCO5zAmFv5nRwF/z+lXzZ+o/wDa5zBY5ht3ya7hMXj4M+BdwK7Bh/hlP6p0qxbgB6fP2Q2BN0t/bH4FJwAKFet4D7srv01rAoo3+fzvHa93oAPzjn4i2GwGkk0cAu5Utfxy4PD8+LP8DaPefT2H77wJTCs+H0kYjAFgk/9HvUrb868CMsmVX5H+YVwDTgH5l+wlgcGHZkvmf7mH5edV/kG3Ef1r+p7xsYdkQUoNjoTbq+XXZP7fLgHsrlJtObgTU+J4MyGW+V1jfJ/9zPbKd96b8JPwT0km2b2HZ53O57QrHPwtYuZ26B+ftNql1/xXWb5jLrFh4De8GVKX8TGBoBz6biwAj8z5eAf5EaogsXSizTV4/oIb6ngR+WHg+Eriq8PxI0km1b34+HLiurI4F8+d/7/z8AeDqsjLntPe5LZSt+BknNV4C+EJhmUgn14vy828BbwOLFcpskLcrNQJ2IZ2AFy2r/5rSsRU+x1uUlRlGasgvWVh2CKkB0Sc/vw4YXrbdornMLoV6PgZWqPW9r/ePbwdYMxiYf99Xtvw+0j8MSFe+T0XES9UqkfTN3K34qqR3gZ/Rsa65dUhXpf+XuwffzfVcDPSTtGyh7HHA/KQGwgERMbNCfQ+VHkTE26SejHUqlOtI/C9H7lItPSf9A10u1zGfpJNz1/YbuZ6jKtTTnlrek5JxpQcR8QnwGvCZDu5vHeDhiPioUNeTpJNrcX+vRsTf26lrY+DtiHis1p3n2xR35NsH/yKdAGH263YF6Yp0qqTfStpHUt9CFecAlyndnjlN0kZt7S8i3o+IPUk9Dz8g9RL9AJgsae12Yl1W0kVKt4Rm5Pd4HeZ8jy8G9pG0ZH7+TWBE4fXdFPhy2ef8TVLPVun20kBSj0jRA3TdOqTG3Kd1RTqj3s/s93ogMDEi3i2UGQf8u1DPpqS/11fLjmPfwjG05YX8d1nyMulveulC/V8rq/tVUkO3WP+LEfFKDftrCDcCrCVI2g+4ELie1PW3IXAGsEAHqin9vexHuuoo/axH+qMvDmBbDfgs6Spjta7EDh2K/6Oy51EW+/+QTibnAV/M8V9G6obvKZVi6qn/Pd0+cFHSysBtpJ6Q/YFNSF3xkF+3fAJaBTiBdLy/BsYpj+SPiJ+Qusz/AKwLPCzpzPb2HRHTI2J4RBxDuuUSwPfb2Ww4sG0uty3pPR7HnO/xX0iNsYOVxqJsDFxaWD8f8Hvm/JxvkI/hsvbirpNoZ/18pGMsP4aBpNtA7Wnvb2k+0mtR6TX6fWG7eg2m7ZT5Gx2AWQ0m5d/bkf4ZU3j+RH48FviGpJWq9AZsBzwREb8sLVDHv1o1idn3JW+rVkjSoqSuwutI/3wvlPRQREwtK7oFqQsZSf1J/+QvrlJtd8Rfquf2iLi8UE/5VdFHpKuZttTynnSnScChkvqWrlYlfR7oB3T0+9xjgSUlbVJjb0DpivI7EfHvvO+Nywvlq9KbgZslnUXqxt+edH+biHgeuAi4SNLJwIlAzfMuRMTbkv5J7tVh9kmq/L3aDvh+RIzMsS5KGofw6esUEbMkXUrqAViTNLaiOFjwMdJYlefyVXglT5HutV9YWLZ1rcfThkmkE+w2zP77EKlBc29h3/tLWjTyN1by52HhsmNYjnSLZgqVVXsNa/EYsH6Fv+um4p4A600Wk7RB2c9aEfEcaVDSRZJ2lrSWpF+TrqjOztteSxqwN1Jp1P0qkgZLGpLXTwbWyyOIV5X0bTr4/fP8T/4s4CxJx0paU2nU9/6Sfl4oeh7pn8pxpCvC+4BrJRWv2gP4haTtJK0HXEm6V35Nld13Of5CPYMk7SBpjXw1unlZmWnAWvnYllGFb1DU+J50pwtIg8GGS1pX0jakq637I+L+tjedy92kruXr8+u5iqStJR1epfwU0vv1P7ns3sCpxQKSTpR0YH7NVgG+AXwCPKv0TYQLJX0hb78h6X71U9UCzLcMzsnv0yqS1pN0Dun1vTkXe4HUbb6b0jdj+uXlk4ED8zYbkP42Kp3kfkcaqHY4cEnZurNIjdKrJG2WY9hB0q8lfS6X+X/AEKVvTqwu6VCgrW881CQiJpEaTpfkv+W1SY2nVfM+AUaQ7rWPyJ+HLYHfku7jlxotfyHdUhipNOfCKpI2kfQdzf6Gyj9JDfudJX0mN8ZrdSawkdI3kjaR9Ln8P+cCNdNkSY0elOAf/0R8OhApKvw8k9cvQbpKfp30h/4YsFNZHcuTTqZvkP6wnyEPxiJ1m19M6rIvjTI/jny7MZcZSjvfDsjlDidd4X9AGpz0CHB0XvfVHN9GhfLLkO7pnl3cD+nbC0/n8o+WbTOIOb8dUEv8pwFTy2KdY/AY6cr5D7mON0lXcT8Bphe2WYp0dT8zb1t6Dacz57cD2nxPmD0wcJuymKYCp7XzGs81MI/Uc3If6b7vjPwaLNfW8bdR/+LA+aSr9Y9IDZ+Tq+2fNKr8xbzvB0gn8QAG5fVHknoY3iEN8BwD7JXXLZRjnZY/M6+Rbuv8Vxvx7ZDLTM/bvEEa6X5gWbnv58/WJ+TBnKTbUw/mWKcDx5BGyA+vsJ+b8+dgwQrr1iMNSHw71zWV1FhYqlDm23n//877OITu+XZAf1Ij5Y382XoE2KGszGb5df6Q1PDZk/SZPbZQZlHSeIwX8vv8T9Jne7uyv+cXSH+Tpf83w0hjDor72zHHu3xh2Yakb/bMyK/BFFJjZIlq9fS2H+VAzaxO8lXIZRHh23HWUJIeBUZHxHfbLdzLSVqD1BjYKSJGNTqeZuF/QmZmLUbSMsAepPki9m9wOJ2SG9PTSFfxnyPdhpoK3NPAsJqOGwFmZq3ndVI3//GRBiw2o2VJYzM+S7qlcT/wPxHRK5KANQvfDjAzM2tR/naAmZlZi/LtAANg5syZ7hIyM5vH9evXT8Xn7gkwMzNrUW4EmJmZtSg3AqzHTJlSbabO3ssx14djrg/HXB/NGHOJGwFmZmYtyo0AMzOzFuVGgJmZWYtyI8DMzKxFuRFgZmbWotwIMDMza1FuBJiZmbUoTxts3apf//6fPt6kgXF0lmOuj2aMmTFjGh2BWbdzT0ATkjRI0lYdKL+BpIckTZI0XtKQnozPzMyag3sCGkySSCmdZ3Vgs0HAu8CDNZZ/H/h6REyR9FlgrKQ7ImJGx6I1M7N5iXsCGkDSAEmTJV0JTAQOzlfqj0u6QdJiudx0Safn5RMkrSVpAHAU8F1J4yRt297+IuLZiJiSH78MvAYs21PHZ2ZmzcE9AY2zOnAIMBW4CdgxIt6TdBLwPeCMXO6NiNhI0jHACRFxuKTfAu9GxDkAkg4ETqywj6kRsW9xgaTNgL7Acz1yVGZm1jTcCGicFyLiYUl7AAOB0enOAH2Bhwrlbsq/xwJfqVRRRFwNXN3eDiWtAPweOKSt2w9dSYbRlAO+zGrUjIliHHN99OaYV1999arr3AhonPfybwGjIuJrVcp9mH9/QpX3q5aeAElLALcCP4yIh9sKrK0PjFkra7a/jSlTpjjmOmjGmEvcCGi8h4ELJa0WEVMlLQqsGBHPtrHNv4AlSk/a6wmQ1Be4GbgyIm7sprjNzKzJeWBgg0XE68BQ4FpJ40m3AtZqZ7M/A1+udWAg8FVgO2Bo3macpA26EreZmTU/9wQ0QERMB9YtPL8b2LRCuQGFx4+RvhpI7iVYvwP7uwq4qrPxmpnZvMmNAOtWM2fMnnqgGe+TOeb6aMaY6cUDv8w6y7cDzMzMWpQbAWZmZi3KtwPMzGqw6a1zDdsxq4sZ3+m5Gd7dE2BmZtai3AgwMzNrUb22ESBpb0kDGx0HgKShki7o5jrnOD5JZ0jasTv3Uba/XXLSoqmSTu6p/ZiZWfPolY0ASfMDe5Pm1O+u+nqbOY4vIk6NiL/2xI4k9QEuBHbN+/xab2lgmZlZ4ygieqbilPL2dlLim42AScDXgROALwELAw8CR0ZESLoXGAdsQ5ri9n+AmflnH+B3wCPADkB/4LCIuD+f4IaRJtJZELgwIi6WNAj4CfA2sFZErFElzoOA40mJex4BjomITyQdCvwAmAE8CXwYEcdJGg7cUpp+V9K7EVFK/XsScBAwC/hLRJws6ZvAEbn+qcDBwAbALWXH96NSvZIGA+eQBm6OAY6OiA8lTQdG5NdvAWC/iHimhvdiS+C0iNg5P/8BQET8rFRm5syZn34QenMiDLNG8cBAa5Qxu4/p0vbFOTn69eun4rqevkJek3SyHi3pcuAY4IKIOANA0u+BPUjT4AL0jYhN8rrVmfNkCzB/RGwmaTfgx8COwGHAzIjYVNKCpGx8d+b6NgLWjYhplYKTtDYwBNg6Iv4j6SLgQEmjgNOBjUkn6XuAJ9o6UEm7AnsBm0fE+5KWyqtuiohLc5kz8+txvqSRFY4PSQsBw4HBEfGspCuBo4Ff5frmSC0MHC5pB+DcCmG9HxFbASsCLxaWvwRsXu1YumsSl2acEMYx10czxmzWKD35t9LTjYAXI2J0fnwV6Yp7mqTvA4sAS5F6CEqNgOvbqa+YVndAfrwTsL6kffPzfsDqwEfAo9UaANlg0ol+TD4JLwy8RjpB3pvn9UfS9UDFnoSCHYErIuJ9gIh4Ky9fN5/8+wOLAXe0U8+awLRCAqERwLHMbgTMlVo4Iu4h9S6YmZnVrKcbAeX3GgK4CNgkIl6UdBqwUGH9e7StUlpdAd+KiDlOrvl2QHv1CRgRET8o23bvNrb5mDyWQtJ8pG7+tgwH9o6IJyUNJc//3wVzvQY19AT8A/ivwvKV8jIzM2thPT0wcOV8PxrgAOCB/PgNSYsB+1beDEjpchevYR93AEdLWgBA0ho5HW8t7gL2lbRc3nYpSf9NGhuwvaSlc737FbaZTuo9ANiTdG8eYBRwqKRFSnXl5YsDr+R6Dqzh+CYDAyStlp8fDPytrYOIiHsiYoMKP1vlImOA1SWtktMK7w+MbKtOMzOb9/V0I2AycKykp4Elgd8AlwITSSfvtkY7XAecKOkJSau2Ue4y4CngcUkTgYupsYcjIp4CTgHuzGl8RwErRMQrwGmktL6jgacLm11KaiA8CWxJ7m2IiNtJJ9bHJI0j3a+HNODvkVxPcRBfxeOLiA+AQ4EbJE0gDTL8bS3H08ZxfgwcR3rNnwb+EBGTulKnmZk1v57+dsAtEbFuO0WtFyh+O6C7NOPgL8dcH465PhxzfTRTzOXfDuiV8wSYmZlZz+uxgYERMR3oFb0AkpYm3f8vNzgi3qx3PGbWfDxPgHWnnkwK1BG9cSa9bpdP9P4KnZmZWYFvB7RB0oONjqGjJN0uaYakW8qWS9JPJT0r6WlJxzcqRjMz6x1aoiegswpfsWsmZ5MmYjqybPlQ0lwBa0XErNLXIs3MrHW5J6ANkt7NvwdJulfSjZKekXS18hSDkjaV9KCkJyU9KmlxSQtJukLShPwVwB1y2aGS/ihplKTpko6T9L1c5uHS3AKSVs1X9GMl3S9prVpjjoi7SHMQlDsaOCMiZuVyr3Xx5TEzsybnnoDabQisA7xM+s7/1pIeJU11PCQixkhaAvg38G0gImK9fAK/U1Jp2uF1c10LkRIKnRQRG0o6l5Rg6VfAJcBRETFF0uakWRa/IOlA4MQKsU2NiLYmXgJYFRgi6cvA68DxEeFMQWZmLcyNgNo9GhEvAeTJgAaQkgu9EhFjACLinbx+G+D8vOwZSS8wO/fAPRHxL+BfkmYyO2/CBFIOhMWArUiTBZX2vWCu62rg6k7GvyDwQURsIukrwOXAtpUKdmcWwWbMSOiY66MZYzbrLvX8/Lc1h4EbAbX7sPC4mLugK/XMKjyfleucD5gREXN9m6GLPQEvMTv50M3AFdUKOougY+5pzRizWXfqLZ9/jwnomsnACpI2BcjjAeYH7ifnCci3AVbOZduVexOmSdovby9Jn8/rrq6SI6C9BgDAH4Ed8uPtgWfbKGtmZi3AjYAuiIiPgCHA+TmXwCjSvf6LgPny3P/XA0Mj4sPqNc3lQOCwXOckYK9aN5R0P3ADMFjSS5J2zquGAfvkmH4GHN6BeMzMbB7k2wFtiIjF8u97gXsLy48rPB4DbFFh80Mr1DeclFq49HxApXURMQ3YpZMxV7zPHxEzgN07U6eZmc2b3AgwM6vBmN3H9Jr7uLVqxrEXjrm+fDvAzMysRbkRYGZm1qJ8O8CsG/Tr379h+96kYXvuvGaMmTFjGh2BWbdzT0AvI2m4pFq+8tfReismFjIzs9blRkCTy/MS1OJs4OCejMXMzJqLGwFdIOlHkiZLekDStZJOqJb8J1/hn5eTDT1futrPkwFdkOv5K7Bcof6NJf0t13WHpBXy8nsl/UrSY6Q8Be1qI7GQmZm1KI8J6KQ8S+A+wOeBBYDHgbFUSf6TN1sB2AZYCxgJ3Ah8GVgTGAh8BngKuFzSAqT8A3tFxOuShgA/Bb6R6+obEZvkWLoynbCZmbUoNwI6b2vgTxHxAfCBpD+TZgusmPwn+2NO5fuUpM/kZdsB10bEJ8DLku7Oy9ckZRwclevqA7xSqOv60oMuJhaaixMIdTzmphzoZh3WKp/nRnPM3csJhOqnavKfrDh1sKqUKa6fFBFbVln/3qcFu7knwAmEmitmq59m+2w04+fZMdeXxwR03mjgS5IWyul/9wDep0rynzbcBwyR1Cff8y8l+ZkMLCtpy1zXApLWqVRBFxMLmZlZi3IjoJNyzoCRwHjgL8AEYCYdT/5zMzCFNBbgSuChXP9HwL7Az3Nd40i3GjqljcRCZmbWonw7oGvOiYjTJC1CuqIfWy35T0QMLXteSk4UwHHl5fO6caQxA+XLB3U00GqJhczMrHW5EdA1l0gaSBoQOCIiHm90QNYYM2fMaNi+m/F+ZDPGTC8e+GXWWW4EdEFEHNDoGMzMzDrLYwLMzMxalHsCzJpYv/79m3KOgmaM2QmEbF7kngAzM7MW5UaAmZlZi2p4I0DS3nmEfcuTdJSkCZLG5aREA/PyL+YkQhPy7y9U2X4pSaMkTcm/l8zLlZMXTZU0XtJG9TwuMzPrnRraCMhpcPcmJc/prvoaTtJSndz0mohYL087/Avgl3n5G8CXImI94BDg91W2Pxm4KyJWB+7KzwF2BVbPP0cAv+lkfGZmNg/pUiNA0gBJz0i6WtLTkm7ME+cg6VRJYyRNlHSJchacsjS4JwF7Amfnq99V8/qfS3pU0rOSts3b9ZF0dq5zvKQj8/JBOWXvSNKse9Vi/WO+ip4k6YjC8sPyfh6VdKmkC/LyZSX9X97fGElbt/NaSNIXJF0DdGoEUUS8U3i6KBB5+RMR8XJePglYWNKC5duTZicckR+PIDWwSsuvjORhoH8pLbGZmbWu7rhyXhM4LCJGS7ocOAY4B7ggIs4AkPR70tz6f87bFNPgrg7cEhE35ucA80fEZpJ2A34M7AgcBsyMiE3zCXC0pDtzfRsB6+bZ+qr5RkS8JWlhYIyk/yNl+PtR3v5fwN3Ak7n8r4FzI+IBSSsDdwBrl1cq6bPAUOBgchrg/Li0/n5g8QrxnBARf61Q37HA94C+zE5BXLQP8HhEfFhh3WciopRp8J+k1MQAKwIvFsq9lJcVsxJ+ylkEmyfmphxl38Sa6bNR4pjrozfH3NNZBF+MiNH58VXA8aRGwA6Svg8sAixFuoItNQKun6uWOd2Uf48FBuTHOwHrSyolxelH6t7+CHi0nQYAwPGSvpwf/1fednngbxHxFoCkG4A1cpkdgYGFlMBLSFosIt4tLZC0GfAgcBmwbUS8Ub7Tjk7XGxEXAhdKOgA4hdT9X9rfOsDPSa9Fe/WEpOjIvkucRbC5Yrb6abbPRjN+nh1zfXVHI6D8RBOSFgIuAjaJiBclnUaaWrfkPdpWusr9pBCjgG9FxB3FgpIGtVdfLrMjsGVEvC/p3rJ4KpkP2CIiPmijzHhSD8VhwJ8kDQeuL3brd7QnoOA6CvfuJa1ESjb09Yh4rso2r0paISJeyd39r+Xl/yA1fEpWysvMzKyFdcfAwJVL6W6BA4AHmH2CfUMpzW5bKW3/ReWTZLk7gKMlLQAgaQ1Ji9YYYz/g7dwAWAvYIi8fA2wvack8qHCfwjZ3At8qPZG0QXmlEfFBRIyIiO1ItwRWBZ6QdFWhzLZV0vxWuhVQbEruTsouiKT+wK3AyYVel0pGMrvn4BDgT4XlX8/jFrYg3VapeCvAzMxaR3c0AiYDx0p6GlgS+E1EzAAuBSaSTt5tDZS7DjhR0hOSVm2j3GWke+6PS5oIXEztPRm3A/PnGIcBDwNExD+As4BHgdHAdFI6YEi3NTbJgxCfAo5qawcRMSUiTiaNkbihxrjKHZcHLo4jjQsondCPA1YDTs0DKMdJWg5A0mWSSreGhwFflDSF1PMxLC+/DXgemEp6X47pZHxmZjYPUcpk28mNpQGkQX3rdldA9Va6z597Am4GLo+ImxsdV73NnDmz8x+EKprxPpljrg/HXB+OuT6aKeZ+/fqp+LzhkwX1AqflK++JwDTgjw2Ox8zMrC66NDAwIqYDvaYXQNLSpElyyg2OiDcrbRMRJ/RsVGZmZr1Tr5hhr7vkE/1cA/jMzLpq01s3bXQITWHGd2Y0OgTrAN8OMDMza1FN3wiQNF3SMvnxg4XlZ+eR9mfnKYAfyd9A6NDkPcX6G0nSWpIekvShpKq3MCStko91qqTrJfXNyxfMz6fm9QPqFbuZmfVOTd8IKIqIrQpPjwDWj4gTgcHAhIjYMCLub0x0iaRFSyfmDnqL2bMxtuXnpOmOVwPeJk1kRP79dl5+bi5nZmYtrEcaAT2ZrEfS0pLuzHVfRppJsLTu3fx7JLAYMFbSSaSMfHvl79cvXKXe30h6LNd7etnq7yul8X1U0mq5/H5KyZGelHRfDa/JZpIuJk2fvGR75ctFxGsRMQb4Txv7ECnfwI15UXkSoVJyoRuBwSrMiWxmZq2npwYG9kiynuzHwAMRcYak3Zl9pfupiNhT0rs5JS+SXiVNYXxcGzH/MMfcB7hL0voRMT6vmxkR60n6OvArUjKkU4GdI+IfeUa/uSilFD4IOJQ0he/lwPGl5D+SridNLlTulxFxZRuxVrM0MCMiPs7PS4mCoJBEKCI+ljQzl58r34ETCDnmemjGmK193fG+NuNnozfH3NMJhCrp9mQ9BdsBXwGIiFslvd1NMX8191rMD6wADCTlBgC4tvD73Px4NDBc0h+YnfDoU0rZBZ8H/gLsGREvlpeJiCHdFHu3cgIhx9zTmjFmq01X39dm/Gw0Y8wl3d4IUM8l6+kxklYBTgA2jYi3lRIBFWOO8scRcZSkzUlz/I+VtHHZXASvknIpHAaMlHQlcHVElJL69ERPwJstlxndAAAgAElEQVRAf0nz596AYqKgUhKhl/LsiP1yeTMza1E9MSagR5L1FNxHOrkiaVc6cX+9giVImQhnSvoMsGvZ+iGF3w/lfa8aEY9ExKnA68yZpY+I+CQiboqI3Um3DxYF7svjJfrlMkOqJBfqTAOASHNA38PshE3lSYRKuQj2Be6OrswZbWZmTa8nbgfcDhyVk/VMppCsR1IpWc9bwDPMmaznQknjc0z3UT1hz+nAtZImAQ8Cf+9qwBHxpKQnckwvkrr6i5bMsX0IfC0vO1sp659IsxQ+SRU5UdGZkn4K7NCZGCUtDzxGarDMkvQdYGBEvCPpNuDwiHgZOAm4TtKZwBPA73IVvwN+L2kq6fXfvzNxmJnZvKPbGwF50Fv5lXTJNRFxSSFZzx/zNm8w+2q7vfrfBHaqsm6xKo+HA8PbqXdoleUD8sOTypZ/pYZwy+sK0oDIDouIf5K69yut263w+HlgswplPgD268y+zQzG7D6m6e77NvO9aquPes8T4GQ9ZmZmvURdcwd0JFmPpEOBb5ctHh0Rx3YlBkmPkL6uWHRwREzoSr1mZmbNptcmEIqIK4AreqDezbu7TjOb9zmBUPucPKj5zFPTBpuZmVnt3AiogaShpSmOu7HOvSUNLDw/Q9KO3bmPsv3tImlyTiB0ck/tx8zMmocbAY2zN2lWQgAi4tSI+GtP7ChPhXwh6VsbA4GvFRsgZmbWmtwIACQdlJMDjZN0saQ+kg4tJTsCti6UHS5p38LzdwuPT8qJhp6UNCwv+2ZOivRkTpK0iKStgD1Jcw2Mk7RqsV5Jg5XSHk+QdLmkBfPy6ZJOl/R4XrdWjYe4GTA1Ip6PiI+A60gJhczMrIX12oGB9SJpbdIcBVtHxH8kXURK+nM6sDFpQqN7SBPvtFXPrqQT6+Z5tsSl8qqbIuLSXOZM4LCIOF8p0+EtEXFjXleqZyHSnAaDI+LZPN3w0aTERQBvRMRGko4hTXV8uKQdmJ3ToOj9nF750+RB2UtA1QGSTiDkmOuhGWO2tnXXe9qMn43eHHMjEgg1k8Gkk/2YfCJeGNgKuDciXodP5/hfo2oNyY7AFRHxPkApURKwbj759yelN76jnXrWBKZFxLP5+QjgWGY3AkrJisYyO5HSPUBbUy13iBMIOeae1owxW/u64z1txs9GM8Zc4kZAmvZ3RET84NMF0t7kE2wFH5Nvo0iaD+jbTv3Dgb3z1MRDgUFdjPfD/PsT8vtXQ09AKXlQSTGxkJmZtSiPCUjz/u8raTmA3I3/BCnZ0dKSFmDO6Xank3oOIN3XXyA/HgUcKmmRQj0AiwOv5HoOLNTzr7yu3GRggKTV8vODgb+1dQARcU+VRERb5SJjgNUlrSKpLylvwMi26jQzs3lfyzcCIuIp4BTgzpwkaBSwAnAaKWPgaODpwiaXkhoITwJbkrIPEhG3k06sj+WpkUuzI/4IeCTX80yhnuuAE/MAwFUL8XwAHArcIGkCMAv4bReP8WPgONKtiKeBP0TEpK7UaWZmzc+3A4CIuB64vmzxw1SYsTAiXmV2emQoJBaKiGHAsLLyvwF+U6Ge0RS+IggMLay7C9iwwjYDCo8fowO3FiLiNuC2Wsubmdm8z40AM7MaOIugzYta/naAmZlZq3JPgJlZDZxAqH1OINR83BNgZmbWotwIMDMza1E91giQ9FlJN7ZT5sH8e4CkA3oqlryPeyVt0k6Z75S+55+f3yapf0/GVYv8+vw75xkYJ6niVwYlLSVplKQp+feSebkknZczCI6XtFF9j8DMzHqjHmsERMTLEbFvO2VKk9kMALrUCMgnuq4ez3eATxsBEbFbRPTITa7SCboDnitMAnRUlTInA3dFxOqkSZBKKYN3BVbPP0dQ4SuLZmbWerrcCJA0TNKxheenSTohX71OzMvWKWTpGy9p9by8lIFvGLBtXv/dauUr7HuApMk5yc5E4L8k7STpoZxp7wZJi1XY7jeSHpM0SdLpednxwGeBeyTdk5dNl7RMtWPMj09UyhI4vlDXopJuVcocOFHSkArhn5iP8UhJS3TwZa9mL1KuAfLvvQvLr4zkYaC/pBW6aZ9mZtakFBFdq0DaEPhVRGyfnz8F7Az0IWXJW1fS+cDDEXF1nra2T0T8W9K7EbGYpEHACRGxR66jYvkK+x4APA9sFREPS1qGlGBn14h4T9JJwIIRcYake/M+HpO0VES8JakP6Yr5+IgYL2k6sElEvJHrnw5sQpp3v9Ixrg3sCxxJykEwEvgFsCywS0R8M5fvFxEzK8S/JvANYB/gfuB3EfFAleOcBDwLvAOcEhH3Vyg3IyL658cC3o6I/pJuAYaV6pZ0F3BSnnAIgJkzZ376QejN2bDMGsXfDmjfmN3HNDoEq6A4V0S/fv1UXNflrwhGxBOSlpP0WdLJ7+2IeDGfuEoeAn4oaSVSat32zjIdKf9CvrqFNJPfQGB0OgfSN9dV7quSjiAd/wp5m/GdOMZvAzsxO83wYqQu9/uB/yfp56SG0Fwn7FzvZOAkSf8LfA24VdKIiDi+rOgrwMoR8aakjYE/SlonIt5pI+aQ1KkWnrMIOuae1owxW/ucRbD5dNc8ATeQroiXZ+7pd4mIayQ9AuwO3CbpyIi4u1plHSz/XuGxgFER8bVqdUtahTSv/6YR8bak4cBCbR8eUPkYBfwsIi6usJ+NgN2AMyXdFRFnVCgjYAdSb8BmwHnAZeXlIuJDcvbAiBgr6TlSauPHyoq+KmmFiHgld/e/lpc7i6CZmc2luwYGXk/KTLcv6WQ5B0mfA56PiPOAPwHrlxWZI6NeDeWreRjYWjkDX743v0ZZmSVIDYeZkj5DGjRXMY4ylY7xDuAbpXEHklYs9Bi8HxFXAWcDc43Gl3QgKaHQscA1wNoR8aOIeKFC2WXzrYvSa7M66TZIuZHAIfnxIaTXrrT863nw5BbAzIh4pcpxmplZi+iWnoCImCRpceAfVU4uXwUOlvQf4J/AWWXrxwOfKGXmGw4s2E75anG8LmkocK2kBfPiU0j30ktlnpT0BOkE/CIpu1/JJcDtkl6OiB3aO8aIuFPS2sBD+fbDu8BBwGrA2ZJmAf8Bjq4Q7gvANhHxeg2Hth1wRn49ZgFHRcRbAJIuA36b7+8PA/4g6bBc/1fz9reReiWmAu+TshSamVmL6/LAQJs3FAcGdpdmvE/mmOvDMdeHY66PZoq5fGCgZww0MzNrUU2RQEjS0qSv8pUbHBFv1jseMzOzeUFTNALyiX6DRsdhZq3L8wS0zRkEm5NvB7QASfvl2RFnqZ38CWZm1jrcCGgNE4GvAPc1OhAzM+s93Aios5zv4BlJwyU9K+lqSTtKGq2U/W8zSdtrdsbAJyQtLmkxSXcp5USYIGmvWvcZEU/n2QnNzMw+1RRjAuZBqwH7kWYKHEPKoLgNsCfwv6S8C8dGxOg8EdEHebsvR8Q7OUfCw5JG5umB76fyJEcnRMRfe/pgzMysObkR0BjTImICgKRJpPS/IWkCKa3ydcAvJV1Nyp3wkqQFgLMkbUeaMGhF4DPAPyNi2+4MrjsTCDVjMiLHXB/NGLNV5/8bvTfmtuYwcCOgMT4sPJ5VeD4LmD8ihkm6lTTL32hJO5OSIy0LbBwR/8kZDhcC6O6eACcQcsw9rRljtrb5/0ZzxVziRkAvJGnV3FMwQdKmwFpAP+C13ADYAfjvUvnu7gkwM7PW4IGBvdN3JE2UNJ6Ue+AvwNXAJvmWwddJuQ9qIunLkl4CtiSlK76jJ4I2M7Pm4p6AOouI6cC6hedDq60r8yHpJN6Zfd4M3NyZbc3MbN7lRoCZWQ3G7D6m6e77NvO9aqsP3w4wMzNrUW4EmJmZtSjfDjAzq4ETCCVOFDRvcU+AmZlZi3IjwMzMrEW5EVBBTu6zbw/Wf5qkEyosHyBpYg/sz6mEzcxsLm4EdANJDRlb0YH9OpWwmZnNpekHBkr6EXAQ8DrwIjCWNDHOhaS59t8HvhkRz0gaDrwDbAIsD3w/Im6UJOB84Iu5jo8K9W8M/BJYDHgDGBoRr0i6FxhHyv53LfD/KsQ2ALgcWCbHd2hE/L2szMa5DMCdheV9gGHAIGBB4MKIuFjSIOAnwNuk6YTXaO81ioinc53tFQWcCMQx10czxmz1ed+a8bPRm2OeZxMI5Xn19wE+DywAPE5qBFwCHBURUyRtDlwEfCFvtgLpxL0WMBK4EfgysCYwkJSZ7yng8py573xgr4h4XdIQ4KekFMAAfSOire7184ERETFC0jeA84C9y8pcARwXEfdJOruw/DBgZkRsKmlBUiKhUiNhI2DdiJiWXwcnEOomjrk+mjFmS3r6fWvGz0YzxlzS1I0AYGvgTxHxAfCBpD+TMuttBdxQuPJdsLDNHyNiFvCUpM/kZdsB10bEJ8DLku7Oy9ckTeM7KtfVB3ilUNf17cS3JakbHuD3wC+KKyX1B/pHxH2FMrvmxzsB6xfGJvQDVif1UjxaagCAEwiZmVnnNHsjoJL5gBkRsUGV9cU0vu31jwuYFBHV5ux/r6PBdYCAb0XEHMl+8u2A98qWdWtPgJmZtYZmHxg4GviSpIUkLQbsQRoDME3SfgBKPt9OPfcBQyT1kbQCsENePhlYVtKWua4FJK3TgfgeBPbPjw8E7i+ujIgZwAxJ2xTKlNwBHJ1vSSBpDUmLVtpJRGwbERtU+HEDwMzMqmrqRkBEjCHd1x9PSrc7AZhJOpkeJulJYBKwVztV3QxMIY0FuBJ4KNf/EbAv8PNc1zjSrYZafQs4NKcEPhj4doUyhwIXShrHnD0Tl+V4Hs9fG7yYTvbcOJWwmZlVoohodAxdImmxiHhX0iKkK/ojIuLxRsfVbGbOnNntH4RmHCzjmOvDMdeHY66PZoq5X79+c9wGnxfGBFwiaSBpQOAINwDMzMxq0/SNgIg4oNExSPohsF/Z4hsi4qeNiMfMul+rJRByoqDW0PSNgN4gn+x9wjczs6bS1AMDzczMrPO6pREgaaikC7qjrkKde+d7/aXnZ0jasTv3Uba/6ZKWyY8fLCw/OyffOVvSspIekfSEpKaaoEfSIZKm5J9DGh2PmZk1Xm++HbA3cAvpa3JExKn12nFEFL8GeASwVER8Iml/YEJEHF5rXZL65JkIG0bSUsCPSTkTAhgraWREvN3IuMzMrLFq6gmQdJCkRyWNk3RxnlTnUEnPSnqUNH1vqewcaXglvVt4fJKkCZKelDQsL/umpDF52f9JWkTSVsCewNl5n6sW65U0OF+NT5B0eZ5bv3Q1f7qkx/O6tdo4pqUl3Zmv8i+j8B39UsySRpISB42VdBJp2t+9ckwLS9pJ0kN5fzfkCYtKcfxc0uPAfjn+2yWNlXR/Ka58TOdJelDS82WvW6XXqmI9NdgZGBURb+UT/yhglxq3NTOzeVS7PQGS1gaGAFtHxH8kXUTK2nc6sDFpcp57gCfaqWdX0qQ9m0fE+/nqFOCmiLg0lzkTOCwizs8n4Fsi4sa8rlTPQsBwYHBEPCvpSuBo4Fe5vjciYiNJxwAnANWu2n8MPBARZ0janZSwZw4Rsaekd0tTEEt6FdgkIo7Ltw5OAXaMiPdyI+F7wBl58zcjYqO83V10IKFRG69VxcRIkg4ETqxwjFMjYl9gRVJ2xJKX8rKKnEXQMddDM8bcShr5/jTjZ6M3x9zVLIKDSSf7MflEvDBp1rx7I+J1AEnX035K2x2BKyLifYCIeCsvXzef/PuTrrrbm81uTWBaRDybn48AjmV2I+Cm/Hsss5P3VLJdaX1E3Cqpo13jW5CyDo7Or0tf8kyD2fWQJjOi4wmN5nqt2qonIq4Gru5g/FU5i6Bj7mnNGHOradT704yfjWaMuaSWRoBIk/D84NMF0t5UP8F+TL7NIGk+0smxLcOBvSPiSUlDgUE1xNSWUoKgT+jZMQ8idbF/rcr6UpKf7kpoVLWeGnoC/sGcr+tKwL1t7MvMzFpALWMC7gL2lbQcfDrI7Alg+3xffQHmnChnOqnnANJ9/QXy41GkefQXKdQDKfvdK7meYgKdf1E5M95kYICk1fLzg4G/1XAc5e4DDsix7Aos2cHtHwa2LsUhaVFJc/WGRMQ7dDyh0VyvVVv1RMTVVRIIlcYY3AHsJGlJSUuS0hQ7f4CZWYtrtxEQEU+R7n3fqZQIZxTpPvZppO7v0cDThU0uJTUQniQlrHkv13M76Z73Y0rJck7I5X8EPJLreaZQz3XAiXkA4KqFeD4gJd25QdIEYBbw244dNpDGNGwnaRKpV+PvHdk43woZClybX5eHSPf1K+lQQqM2XquOJkYq1fcW8BNgTP45o3A7xszMWlTTJxCy7uEEQoljrg/HXB+OuT6aKebyBEKeMdDMzKxF9ebJgrqFpEOBb5ctHh0RxzYiHjMzs95inm8ERMQVwBWNjsPMmlsrZRF0BsHW4dsBPUTSvZI2yY9vk9Q//xxTKPNZSTd2sv45ZmZsp+yBksbnGQgfrOHbCWZm1gLcCKiDiNgtImaQJkQ6prD85cLX+HrSNGD7iFiP9C2BS+qwTzMz6+XcCKiRpAGSnpF0taSnJd2olOegYh6Dsm1LGQqHAasq5R44O9c5MZfpI+kcSRPzVfu38vJTlXIrTJR0iQrTBdYqIh4sJAt6mDRZkJmZtTg3AjpmTeCiiFgbeIeUK2A4MCRfZc9PymNQzcnAc3kin/IZ/o4ABgAbRMT6zJ4G+IKI2DQi1iVN2bxHeaWSzs0Ni/KfkyvEcBjwl1oP2MzM5l3z/MDAbvZiRIzOj68iTXTUVh6DjtgR+G1EfAxz5FbYQdL3gUWApUiTBP25uGFEfLeWHUjagdQI2Katck4g5JjroRljbhWNfm8avf/O6M0xdzWBkM1WPqHODGDpntpZzph4ESlz4YuSTgMWqlDuXGCHClVcFxGlNMTrA5cBu0bEm23t1wmEHHNPa8aYW0kj35tm/Gw0Y8wlbgR0zMqStoyIh0h5Bx4DjpS0WkRMpf08BtXyIUCajvlISfdExMc5t8KsvO6NnEVwX2CubxO01xMgaWVSdsWDC70WZmbW4jwmoGMmA8dKepqUcOhcOpDHIF+Bj86D/M4uW30ZKX/B+Jwb4ID8jYJLgYmkhD9jOhn3qaQei4vyWIHHOlmPmZnNQ9wT0DEfR8RBZcvuAjYsLxgRgwqPBxQeH1BWdN28/GPSQMPvldVzCimBU3n9Q2sNOiIOBw6vtbyZmbUGNwLMzGowZvcxTXfft5nvVVt9uBFQo4iYTr5qNzMzmxd4TICZmVmLck+AmVkNWiGBkBMHtR73BJiZmbUoNwJ6SKUsgu2UP0PSjj0Yz3clTcpfT7w2T0RkZmYtzI2ALlDS7mtYyCLYVplTI+Kv3RfdbJJWBI4nzTy4LtAH2L8n9mVmZs3DjYAOypn/Jku6kjSJz8GSHpL0uKQb8sx+5duUsggi6Ud5+wfyFfkJeflwSfvmxxUzE+Z6Ts/7miBprQ6EPj+wsKT5SXkIXu7aK2FmZs3OAwM7Z3XgEGAqaTreHSPiPUknkSb7OaPSRpI2BfYBPg8sADwOjC0rsxApM+HgiHg2NzaOZnZSojciYiNJxwAnAIfnxEDnVtjl+xGxVUT8Q9I5pBkJ/w3cGRF3Vjs4JxByzPXQjDHP63rLe9Jb4uiI3hyzEwh1vxci4mFJewADSVMBA/QFHmpju62BP0XEB8AHkv5cocyatJ2Z8Kb8eyzwFYCIuAfYoNpOJS0J7AWsQkp6dIOkgyLiqkrlnUDIMfe0Zoy5FfSG96QZPxvNGHOJGwGd817+LWBURHytjvv+MP/+hPz+tdcTQEpTPC0iXs/lbwK2IqVDNjOzFuUxAV3zMLC1pNUAJC0qaY02yo8GviRpoTx2YI8KZSYDA0p10n5mQiLinojYoMLPVrnI34EtJC2i1GUxGHi69sM0M7N5kRsBXZCvrIcC10oaT7oVUHWwXkSMAUYC44G/ABOAmWVlPqADmQlrjPMRUgrix/M+5wMu6UqdZmbW/Hw7oIPKcwhExN3AXFOJVcsiCJwTEadJWgS4jzwwsJgVMCKqZSYcUHj8GDCovEwbcf8Y+HGt5c3MbN7nRkD9XSJpILAQMCIiHm90QGbWPmcRtHmRGwF1FhEHNDoGMzMzcCPAzKwmTiBk8yIPDDQzM2tRbgR0ozyl8MRGx1FO0oKSrpc0VdIjkgY0OiYzM2s8NwJaw2HA2xGxGmlSoZ83OB4zM+sF3Ajofn0kXZrT9t4paWFJG0h6WNJ4STfnaXzL0w0vI2l6fryOpEcljcvbrJ6XH1RYfrGkPjXGtBdp+mFI8wUMzpMGmZlZC3MjoPutDlwYEeuQ5unfB7gSOCki1idN1tPe9/WPAn4dERsAmwAvSVobGAJsnZd/AhwIkLv6x1X4+Xqub0XgRYCI+Jg0QdHS3XjMZmbWhPztgO43LSLG5cdjgVWB/hFRmvp3BHBDO3U8BPxQ0krATRExRdJgYGNgTL6IXxh4DSAihnTnATiLoGOuh2aMeV7XW96T3hJHR/TmmJ1FsL4+LDz+BOjfRtmPmd0bs1BpYURcI+kRYHfgNklHkpIVjYiIH5RXIul6UvbBcr+MiCuBfwD/RepRmB/oB7xZLShnEXTMPa0ZY24FveE9acbPRjPGXOJGQM+bCbwtaduIuJ85EwJNJ13dPwrsW9pA0ueA5yPiPEkrA+sDdwJ/knRuRLwmaSlg8Yh4oYaegJHAIaQehn2BuyMiuu8QzcysGbkRUB+HAL/N+QKeJyUIAjgH+IOkI4BbC+W/Chws6T/AP4GzIuItSacAd0qaD/gPcCzwQg37/x3we0lTgbeA/bvjoMzMrLm5EdCNKiQXOqeweosK5Z8hXeWXnJKXDwOGVSh/PXB9J+L6ANivo9uZmdm8zY0AM7MaOIGQzYv8FUEzM7MW5UaAmZlZi/LtADOzGvSGLILO8mfdzT0BZmZmLaolGwGSPivpxvx4A0m71bDNIEm3dHA/AyQd0F3lOkvS0pLukfSupAt6aj9mZtZcWq4RIGn+iHg5IkqT82wAtNsI6KQBQC0n91rLddYHwI+AE3pwH2Zm1mSaphGQr5afkTRc0rOSrpa0o6TRkqZI2iz/PCTpCUkPSlozbztU0khJdwN35bomSuoLnAEMyQl3hlSro4b4ti8k7nlC0uKk7/pvm5d9N+/3fkmP55+t8ubl5YYWr9gl3ZJ7Ivrk458oaYKk79YSW0S8FxEPkBoDZmZmQPMNDFyNNOnNN4AxpKvnbYA9gf8Fvg5sGxEfS9oROIuUxQ9gI2D9PPPeAICI+EjSqcAmEXEcgKQl2qijLScAx0bEaEmLkU64JwMnRMQeue5FgC9GxAc5PfC1pCyB5eWGVtnHBsCKEbFuLtc//z6RnFGwzH0RcXwNsc/BCYQccz00Y8yN1pnXrBlfZ8fcvealBELTImICgKRJwF0REZImkLrU+wEj8gk2gAUK246KiLdq2EdbdbRlNPBLSVeTMv+9lLP9FS0AXCCplAp4jRrrLnke+Jyk80nTDN8JEBFnA2d3sK6qnEDIMfe0Zoy5N+joa9aMr7Njrq+muR2QFTP0zSo8n0Vq0PwEuCdfKX+JQmY+4L0a99FWHVXlqX4PJ6X4HS1prQrFvgu8Cnye1APQt0p1xeyClGKIiLfztvcCRwGXQeoJKNyKKP6cV0vsZmbWmpqtJ6A9/UhpcwGG1rjNv4DFu1gHklbNvRQTJG0KrAW8WKHulyJilqRDgD5VYpgOHJMTBa0IbJb3sQzwUUT8n6TJwFXQ/T0BZmbWGpqtJ6A9vwB+JukJam/g3AMMLA0M7GQdAN/JA/bGkzL8/QUYD3wi6ck8iO8i4BBJT5IaCaXeifJyo4FpwFPAecDjudyKwL2SxpEaAD+oNThJ04FfAkMlvSRpYAeOzczM5kFyWnkDmDlzZrd/EJrxPpljrg/HXB+OuT6aKeZ+/frNMVhtXusJMDMzsxrNa2MCepykQ4Fvly0eHRHHNiIeMzOzznIjoIMi4grgikbHYWZm1lW+HWBmZtai3AgwMzNrUW4EmJmZtSg3AszMzFqUGwFmZmYtypMFGdAzkwWZmVnv4smCzMzMDHAjwMzMrGX5doCZmVmLck+AmZlZi3IjwLpE0lKSRkmakn8vWaXcIbnMFEmHFJZ/TdIESeMl3S5pmSaIua+kSyQ9K+kZSfv09pgL60dKmtjT8eZ9dTpmSYtIujW/vpMkDevhWHeRNFnSVEknV1i/oKTr8/pHJA0orPtBXj5Z0s49GWd3xCzpi5LG5r+7sZK+0NtjLqxfWdK7kk5ohpglrS/pofwZniBpoXrFXbOI8I9/Ov0D/AI4OT8+Gfh5hTJLAc/n30vmx0uScle8BixTqOu03hxzXnc6cGZ+PF8p/t4cc17/FeAaYGITfDYWAXbIZfoC9wO79lCcfYDngM/lfT0JDCwrcwzw2/x4f+D6/HhgLr8gsEqup08dXtuuxLwh8Nn8eF3gH3X6PHQ65sL6G4EbgBN6e8yk/2/jgc/n50vX47PR0R/3BFhX7QWMyI9HAHtXKLMzMCoi3oqIt4FRwC6A8s+ikgQsAbzc8yF3KWaAbwA/A4iIWRHxRg/HC12MWdJiwPeAM+sQa0mnY46I9yPiHoCI+Ah4HFiph+LcDJgaEc/nfV2XYy8qHsuNwOD8md0LuC4iPoyIacDUXF9P63TMEfFERJT+ziYBC0tasDfHDCBpb2BajrleuhLzTsD4iP/f3v2EWlHGYRz/PnALd1mRpt7qWthCCVSUCIqEykVQm1y0KKNsYbtykch1IRpBUiQVBFHrIKHoRoHgDRFqkSj2h8I/idC9WAsXQUWR+XPxvgeH0zHPPXPe6Vzm+cDAmTnvDI/DzJnfzPvONb4GiIjzEfFPQ7n75iLA6locEefy55+BxT3aLAN+qszPAMsi4m/gOeBb0sV/JfBewawdA2eWtDDP75F0TNJ+Sb3WH7aBM7mNEfwAAALYSURBVOfPe4DXgD+KJfy3upkByPv8EWC6RMh+MlTbRMQF4FfSnV0/65ZQJ3PVY8CxiPirUM6eebK+M+cidjvpKVyT6uznO4GQdCD/VrzYQN45838lbFcl6SBwc4+vJqszERGS+n7dRNI1pCJgDekx8JvADoZwt1oqM+mcGQe+jIhtkrYBrwJPDhw2K7ifVwN3RMQL3X2sdRXcz53tjwHvA29ExJnBUlovklYBr5DuWEfdLuD1iPgtPxiYD8aAe4H1pOJ7WtLRiChVzA7ERYBdVUQ8eKXvJP0iaUlEnJO0hNTH320W2FCZHwcOAavz9n/M2/qA1Hc8ypnPk07oD/Py/cCWEc98D7BO0lnSOb9I0qGI2EBNBTN3vAOcioh9dbP+h1nglq4Ms1doM5MLk+tIx0I/65ZQJzOSxoGPgM2d868BdTLfDWyStBdYCFyU9GdEvDXCmWeAw53uQkmfAWsp90RrIO4OsLqmgM4o9KeAj3u0OQBslHR9HiG+MS+bBVZKuim3ewj4oXBeqJE50gifT7h84XoA+L5sXKBe5rcjYmlETJDuTE4OowAomRlA0kukH9TnC+c8AqyQtFzStaTBXVNdbar/lk3A5/lYmAIezyPElwMrgK8K562VOXevfEoatPlFA1k7Bs4cEfdFxEQ+hvcBLzdQANTKTDqO71J602UMuJ9mfivm5v8emehpfk+kvq9p4BRwELghL18HvFtp9wxp0NRp4OnK8q2kC/83pIvrjfMg823A4Zx5Grh11DNXvp+gubcDBs5MuuOKfGwcz9OzBbM+DJwkjQSfzMt2A4/mzwtIT31Oky7yt1fWnczrnaDQGwzDzAzsBH6v7NfjwKJRzty1jV009HbAEI6NJ0gDGb8D9jaVeS6T/2KgmZlZS7k7wMzMrKVcBJiZmbWUiwAzM7OWchFgZmbWUi4CzMzMWspFgJmZWUu5CDAzM2spFwFmZmYtdQmpT8NScFFu2AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56a2929a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "# Ensemble Probability of Relationship Success = 83.5%"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgEAAAEGCAYAAAD8C0ZEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzs3Xm8XdP9//HXW8xTYqbUL0oMMRYxDwmqhra0aLSmkLa0VPGl9NtWTVXKt4aWmkpiKEppldYsFTHFkIQgEhI11Zy0KIrP74+1jrvvcc695w7n3nvueT8fj/u45+y99tprrzPstddeZ30UEZiZmVnzmau3C2BmZma9w40AMzOzJuVGgJmZWZNyI8DMzKxJuRFgZmbWpNwIMDMza1JuBJhlkkZJ+rC3ywEgabikkLRCb5elRNIsST/poX0Nl/S4pP9KGtcT+yzbf0jau6f32+j64vu2MySdIunx3i5HT3AjwPoESWMk3d7b5TCQdFGVE+8w4IweKsZvgUeAzwFf66F99imSviLpHklvSnpH0gxJV0haNK/fIp9wB/duSWuXy9vW36weLs8qeb+b9OR++xI3AsysJhHxWkS800O7GwLcFhHPR8SbnclA0lySBnRzuXqEpG2A64DbgC2AdYCDgX8B8/Vi0bpqucLfbnnZ+oVlw3qpXD1C0gBJfeq826cKY1aNpEUknS/pNUnvS3pI0vZlaZaWdImkVyS9J2mapAPyOkm6UNIzkv4j6VlJJ0vq8BeqpO9LeirvY7qkH0uaO69bRdK/JB1eSL9GvpL7Tn4+StKHkraTNDXn84Ck9drYZ7vll3RcvlrcJZfvHUnjJA0ppFlM0uWS/pHzmSbpfySplAcwGti6cHU2Kq9rdTugvddE0uC8/dcl3Sjp3VzuUW0c53BJAQwALi3b/yaS7s7lfkvS7yUtXeH4R0p6CvgAWLXKfhaWdKak53PZZ0n63zbK9QNJkyS9Lemfkq6StFxh/TySfiXphZzfy5KuKqxfU9Itkmbn1+VJSftU2x/wFWBSRBwfEU9ExDMRcUtEfDciXlO6+h+f087M9TQu72t9SX+T9Gou70RJO5TV07QKx3ixpDsKzzeQdGvO4zVJ10n6f2XbfD8f87uSbgFWbOOYiIh/lv6AUuPutcLy13K+gyT9TtLrhc/HiLJ9D8vH9n5+H++SX5sjC2kWlXROfj3eye/RL+d18wPTc9L7ch0+VbaPPSQ9nevgDkkrla3fOC9/R+l75w8q3ApRvq0gaW9JTwPvA63qsNdFhP/81+t/wBjg9jbWXwPMAr4IrAGcRfqSXz2vXwB4ktSFvB2pG3l7YM+8fi7g58DGwGDSl+zLwPGFfYwCPmynnMcBzwFfBVYCdgL+AZxYSLMX6cO+PjA/MAX4Q9l+Ps5l3Zp0lXcj8CKwQE4zHAhghQ6U/zjgHeBmYANgXeBhYHwhzbLAMblsKwF7A28D++f1CwNXAPfmtMsWyjQL+EkHXpPB+RieBb4OrAKcDHwIrFqlfufN+wzSle+y+bVdlnQV/HtgbdLV8RTg7rLjfxf4e66nVYFFKuxDwLhcrl1J75WtgG8X0gSwd+H5D0jvq5WATXP9/L2w/gjghfy6rUi6oj2ssH5KLvvQvL8dgS+18T47GpgNbFRl/YD8Hoi8r2WBxQvvnVHAmrkOTsqvy6p5/Qr5Ndi6kN8i+X0wMj8fmp8fD6ye6/wa4Glg/pxml5zPEXk/o4FXKLxv2/ksDa+WFvgL8Eyu86Gk20PvAZ/L6xcFXiP1lqwNbAY8kNMcWfjMTABuz+tXzu+p/wJb5DSb5DLslOtwybz8FODfwE3A5/PfFFLvVKmM65Hebz8GViN93v4ETAXmKeTzDnBHfp1WBxbq7e/bVnXd2wXwn/8i2m4EkE4eAexUtvwR4OL8eHT+Amj3y6ew/eHA9MLzUbTRCAAWzB/6HcqW7wvMLlt2Sf7CvASYCQws208A2xaWLZa/dEfn51W/INso/3H5S3mpwrKRpAbH/G3kc1bZl9tFwLgK6WaRGwE1viaDc5ojCusH5C/XA9t5bcpPwieSTrLzFpatm9NtVTj+j4EV28l727zdhrXuv8L6z+c0yxfq8E5AVdLPAUZ14L25IHBD3sfLwJ9JDZElCmm2yOsH15DfZODHhec3AJcXnh9IOqnOm5+PAa4qy2O+/P7fNT+/B7iiLM3p7b1vC2krvsdJjZcAtiksE+nkem5+/n3gLWDhQpr18nalRsAOpBPwQmX5/750bIX38SZlaU4hNeQXKyzbj9SAGJCfXwWMKdtuoZxmh0I+HwLL1fra9/SfbwdYIxia/99dtvxu0hcGpCvfJyLihWqZSPp27lZ8RdLbwC/oWNfcmqSr0j/m7sG3cz7nAwMlLVVIewgwN6mB8M2ImFMhv/tKDyLiLVJPxpoV0nWk/C9F7lItPSd9gS6d85hL0jG5a/v1nM9BFfJpTy2vScmk0oOI+Ah4FVimg/tbE7g/Ij4o5DWZdHIt7u+ViPhHO3ltALwVEQ/VuvN8m+KWfPvg36QTILTU2yWkK9IZks6TtJukeQtZnA5cpHR75jhJ67e1v4h4NyK+Qup5+BGpl+hHwDRJa7RT1qUknat0S2h2fo3XpPVrfD6wm6TF8vNvA2ML9TsM+GrZ+/wNUs9W6fbSUFKPSNE9dN2apMbcJ3lFOqOOp+W1Hgo8HhFvF9JMAv5TyGcY6fP6Stlx7F44hrY8lz+XJS+RPtNLFPL/Rlner5AausX8n4+Il2vYX69wI8CagqQ9gHOAq0ldf58HTgDm6UA2pc/LHqSrjtLf2qQPfXEA2yrAZ0hXGat0pezQofJ/UPY8ysr+P6STydnAF3L5LyJ1w9dLpTLV67un2wcuSloR+CupJ2RPYENSVzzkessnoJWAI0nHexYwSXkkf0ScSOoy/wOwFnC/pJPa23dEzIqIMRHxPdItlwB+2M5mY4Atc7otSa/xJFq/xn8jNcb2URqLsgFwYWH9XMBltH6fr5eP4aL2yt1Dop31c5GOsfwYhpJuA7Wnvc/SXKS6qFRHlxW266nBtJ0yd28XwKwGU/P/rUhfxhSeP5ofPwwcIGmFKr0BWwGPRsSvSgvU8Z9WTaXlvuRfqyWStBCpq/Aq0pfvOZLui4gZZUk3IXUhI2kQ6Uv+/CrZdkf5S/ncHBEXF/Ipvyr6gHQ105ZaXpPuNBXYX9K8patVSesCA4GO/p77YWAxSRvW2BtQuqI8LCL+k/e9QXmifFV6PXC9pJNJ3fhbk+5vExHPAucC50o6BjgKqHnehYh4S9I/yb06tJykyl+rrYAfRsQNuawLkcYhfFJPEfGxpAtJPQCrkcZWFAcLPkQaq/JMvgqv5AnSvfZzCss2r/V42jCVdILdgpbPh0gNmnGFfe8paaHIv1jJ74cFyo5hadItmulUVq0Oa/EQsE6Fz3VDcU+A9SULS1qv7G/1iHiGNCjpXElflLS6pLNIV1Sn5W2vJA3Yu0Fp1P1KkraVNDKvnwasnUcQryzpB3Tw9+f5S/5k4GRJB0taTWnU956STi0kPZv0pXII6YrwbuBKScWr9gB+KWkrSWsDl5Lulf++yu67XP5CPsMljZC0ar4a3bgszUxg9XxsS6rCLyhqfE26029Ig8HGSFpL0hakq63xETG+7U0/5U5S1/LVuT5XkrS5pG9VST+d9Hr9T067K3BsMYGkoyTtletsJeAA4CPgaaVfIpwjaZu8/edJ96ufqFbAfMvg9Pw6rSRpbUmnk+r3+pzsOVK3+U5Kv4wZmJdPA/bK26xH+mxUOsn9jjRQ7VvABWXrTiY1Si+XtFEuwwhJZ0n6XE7zf8BIpV9ODJG0P9DWLx5qEhFTSQ2nC/JneQ1S42nlvE+AsaR77WPz+2FT4DzSffxSo+VvpFsKNyjNubCSpA0lHaaWX6j8k9Sw/6KkZXJjvFYnAesr/SJpQ0mfy985v1EjTZbU24MS/Oe/iE8GIkWFv6fy+kVJV8mvkT7oDwHbl+WxLOlk+jrpg/0UeTAWqdv8fFKXfWmU+SHk2405zSja+XVATvct0hX+e6TBSQ8A383rvp7Lt34h/ZKke7qnFfdD+vXCkzn9g2XbDKf1rwNqKf9xwIyysrYaPEa6cv5DzuMN0lXcicCswjaLk67u5+RtS3U4i9a/DmjzNaFlYOAWZWWaARzXTh1/amAeqefkbtJ939m5DpZu6/jbyH8R4Nekq/UPSA2fY6rtnzSq/Pm873tIJ/EAhuf1B5J6GP5FGuA5Edglr5s/l3Vmfs+8Srqt89k2yjcip5mVt3mdNNJ9r7J0P8zvrY/IgzlJt6fuzWWdBXyPNEJ+TIX9XJ/fB/NVWLc2aUDiWzmvGaTGwuKFND/I+/9P3sd+dM+vAwaRGimv5/fWA8CIsjQb5Xp+n9Tw+QrpPXtwIc1CpPEYz+XX+Z+k9/ZWZZ/n50ifydL3zSmkMQfF/W2Xy7tsYdnnSb/smZ3rYDqpMbJotXz62p9yQc2sh+SrkIsiwrfjrFdJehCYEBGHt5u4j5O0KqkxsH1E3Nbb5WkU/hIyM2sykpYEvkSaL2LPXi5Op+TG9EzSVfznSLehZgB39WKxGo4bAWZmzec1Ujf/oZEGLDaipUhjMz5DuqUxHvifiOgTQcAahW8HmJmZNSn/OsDMzKxJ+XaAATBnzhx3CZmZ9XMDBw5U8bl7AszMzJqUGwFmZmZNyo0Aa2jTp1ebDbQ5uT5ac3205vpo4bpI3AgwMzNrUm4EmJmZNSk3AszMzJqUGwFmZmZNyo0AMzOzJuVGgJmZWZNyI8DMzKxJedpgM+u3ht00rLeLYNZlsw+bXbe83RNgZmbWpPpsI0DSrpKG9nY5ACSNkvSbbs6z1fFJOkHSdt25j7L97SBpmqQZko6p137MzKxx9MlGgKS5gV2BbmkE5Pz6mlbHFxHHRsTt9diRpAHAOcCOeZ/f6CsNLDMz6z11awRIGizpKUlXSHpS0rWSFpR0rKSJkh6XdIEk5fTjJJ0p6SHgaOArwGmSJklaOa8/VdKDkp6WtGXeboCk03KeUyQdmJcPlzRe0g3AE22Uc++c5yRJ5+cTJpL2z/t5ENi8kH6MpN0Lz98uPD5a0mOSJks6JS/7di7bZEl/zHWwWYXj+yRfSdtKejTndbGk+fLyWZKOl/RIXrd6jS/HRsCMiHg2Ij4ArgJ2qXFbMzPrp+p9hbwaMDoiJki6GPge8JuIOAFA0mXAl4C/5PTzRsSGed0Q4MaIuDY/B5g7IjaStBPwM2A7YDQwJyKG5ZPlBEm35vzWB9aKiJmVCidpDWAksHlE/FfSucBekm4Djgc2AOYAdwGPtnWgknYknVg3joh3JS2eV10XERfmNCfl+vh1bpyUHx+S5gfGANtGxNOSLgW+C5yZ83s9ItaX9D3gSOBbkkYAZ1Qo1rsRsRmwPPB8YfkLwMbVjqXRAms0WnnrzfVh1r909TM9ZMiQquvq3Qh4PiIm5MeXA4cCMyX9EFgQWByYSksj4Op28rsu/38YGJwfbw+sU7g6HwgMAT4AHqzWAMi2JZ3oJ+aT8ALAq6QT5LiIeA1A0tXAqu2UbTvgkoh4FyAi3szL18on/0HAwsAt7eSzGjAzIp7Oz8cCB9PSCCjWwdfyvu4C1msn35q19Ybpa6ZPn95Q5a0314dZ/1PPz3S9GwFR4fm5wIYR8byk44D5C+vfaSe/9/P/j2gpu4DvR0Srk6uk4TXkJ2BsRPyobNtd29jmQ/JtFElzAfO2s48xwK4RMVnSKGB4O+nb86k6qKEn4EXgs4XlK+RlZmbWxOo9MHBFSZvmx98E7smPX5e0MLB75c0A+DewSA37uAX4rqR5ACStKmmhGst3B7C7pKXztotL+n/AA8DWkpbI+e5R2GYWqfcA0n39efLj24D9JS1YyisvXwR4OeezVw3HNw0YLGmV/Hwf4O9tHURE3BUR61X42ywnmQgMkbSSpHmBPYEb2srTzMz6v3o3AqYBB0t6ElgM+C1wIfA46eQ9sY1trwKOygPkVm4j3UWkgX+PSHocOJ8aezgi4gngJ8CtkqaQTuTLRcTLwHHAfcAE4MnCZheSGgiTgU3JvQ0RcTPpxPqQpEmk+/UAPyU1KiYAT7V3fBHxHrA/cI2kx4CPgfNqOZ42jvND4BBSnT8J/CEipnYlTzMza3yKKO+x76aMpcGkgW9r1WUH1q3mzJlTnzdCnfkeeGuuj9ZcH625Plo0a10MHDhQxed9cp4AMzMzq7+6DQyMiFlAn+gFkLQE6f5/uW0j4o2eLo+ZmVlf0Bdn0ut2+UTfbT+hs+oGDhrUo/vbsEf31ve5PspMbGvYkZn5doCZmVmTciPAzMysSfV6I0B9KFpgb5N0UI4JMEnSPaV6kfQFSQ/ndQ9L2qbK9otLuk3S9Px/sbxcks5WiiA4RdL6PXlcZmbWN/VqI0D9NFpgYaKgjvp9RKwdEesBvwR+lZe/Dnw5ItYG9gMuq7L9McAdETGENBCyFDJ4R9JUykOA75DmazAzsybXpZNmngvgZtI89uuT4gDsmwPoHAt8mTQf/73AgRERksYBk4AtgOtJs+5tLeknwG7A70iT64wgzbc/OiLG5+h+p5Cm3Z0POCcizs/TA58IvAWsTpU5/iX9iTR17vzAWRFxQV4+mhS1cDYwGXg/Ig6RtBRpkp4VcxaHFeIgVMpfuczfIsUeaGuCo4oi4l+FpwuRp12OiGLwoqnAApLmi4j3aW0XWqYlHguMy8e2C3BppEkh7pc0SFJpUqRP6UqwCg9Ms77GAZVac320aJa6qHcAoUqRAk+nD0ULzA6IiDclLUAKGPRHUmPip3n7fwN3khoCAGcBZ0TEPZJWJM22t0Z5ppI+A4wiTe/7BHBxflxaP57K0wMfGRG3V8jvYOAIUkyCSt3+uwGPVGgAACxTOLH/E1gmP64URXB5oGIjoBkn0LD+y+/nFs06QU4lroukOxoBlSIFng6M6EPRAgEOlfTV/Pizedtlgb+XIv5JuoaWnoTtgKGlEL/AopIWjoi3SwskbUTq5bgI2DIiXi/faURs2U65ytOfA5wj6ZukKY33K+xvTeBUUl20l09IashZAM3MrGd0RyPgU5ECJc1PH4oWmNNsB2yab1WMKytPJXMBm+S5/KuZQuqhGA38WdIY4Opit35HewIKrqJw717SCqTbJ/tGxDNVtnml1M0vaTlSWGRwFEEzM6ugOwYGVooUWDrB9pVogQOBt3IDYHVgk7x8Imk8wmJ5UOFuhW1uBb5feiLpU5MNRcR7ETE2IrYi3RJYGXhU0uWFNFtWifBX6VZAsW9qZ2B6Xj4IuAk4pq1xCaQARqWeg/2APxeW75t/JbAJ6bZKxVsBZmbWPLqjEfCpSIERMZs+FC2QNHhx7lzGU4D7ASLiReBk4EFSlL9ZwJy8zaHAhvkndU8AB7W1g4iYHhHHkMZIXFNjucodImlqjkJ4BC0n9EOAVYBj888HJ6kl/PFFkkrj8U4BviBpOqnn45S8/K/As8AM0uvyvU6Wz8zM+pEuRRHsD5ECS/f5c0/A9cDFEXF9b5erpzmKYP/g+mjN9dGa66NFs9aFowh+2nH5yvtxYCbwp14uj5mZWY/o0sDAvhQpEDoXLTAijqxvqcystwy7aVhvF8GM2YfN7u0iVNUnZtjrLo4WaGZmVjvfDmiDpHt7uwwdJelmSbMl3Vi2XJJ+LulpSU9KOrS3ymhmZn1Dv+oJ6G4RsVlvl6ETTiNN0HRg2fJRpLkCVo+Ij0u/LjAzs+blnoA2SHo7/x8uaZykayU9JemKHCsAScMk3StpsqQHJS0iaX5JlyhF/XtU0oicdpSkPylF+Jsl6RBJR+Q095cCD0laOV/RPyxpfJ7boCYRcQdp7oVy3wVOiIiPc7pXK6QxM7Mm4p6A2n0eWBN4iTSnwOaSHiRNgTwyIiZKWhT4D/AD0sy9a+cT+K2SStMRr5Xzmp/0u/2jI+Lzks4A9gXOBC4ADoqI6ZI2Js2+uI2kvYCjKpRtRkS0NSETpImMRuapk18DDo2I5oieYWZmFbkRULsHI+IFgPyTwsGkiYVejoiJ0BIFUNIWwK/zsqckPUdLTIK7IuLfwL8lzaElnsJjpNgICwObAdcU4hbMl/O6Ariik+WfD3gvIjaU9DVSoKOKcQ0aLbJWo5W33lwfZn1Lb38m6x1FsFkUo/YVYxp0JZ+PC88/znnOBcyOiE/9yqGLPQEv0BKY6XrgkmoJG2kCjWad8KMa14dZ39OXP5MeE9A104DlJA0DyOMB5gbGA3vlZasCK+a07cq9CTMl7ZG3l6R187orqsQhaK8BAGkSpBH58dbA07UfppmZ9UduBHRBRHwAjAR+LWkycBvpXv+5wFySHiONGRgVEe9Xz+lT9gJG5zynArvUumGOWngNsK2kFyR9Ma86Bdgtl+kXwLc6UB4zM+uHuhQ7wPoPxw7oH1wfrQ06c1BvF8GsT80YWB47wGMCzKzfmrjzRDeKCtxIbOG6SHw7wMzMrEm5EWBmZtakfDvAzPotRxFsTn3pHnxf554AMzOzJlW3RoCkz0i6tp009+b/gyV9s15lyfsYJ2nDdtIcJmnBwvO/Sur14cW5fv4jaVL+O69KusVzXILp+f9iebkknS1phqQpktbv2SMwM7O+qG6NgIh4qb1JbApR+gYDXWoE5BNdV4/nMFIEPgAiYqeIqEu/UukE3QHPFCYHOqhKmmOAOyJiCHBHfg6wIzAk/30H+G1nymxmZv1LlxsBkk6RdHDh+XGSjsxXr4/nZWvmCHuT8pXokLz87bzZKcCWef3h1dJX2PdgSdMkXQo8DnxW0vaS7pP0iKRr8lz85dv9VtJDkqZKOj4vOxT4DHCXpLvyslmSlqx2jPnxUZIm5nKW8lpI0k1KkQUflzSyQvGPysd4YA481B12Acbmx2OBXQvLL43kfmCQpOW6aZ9mZtagumNg4NWkyHfn5OdfB74IDCikOQg4KyKukDRv2TpIV6xHRsSXACT9up30RUOA/SLifklLAj8BtouIdyQdDRwBnFC2zY8j4k1JA4A7JK0TEWdLOgIYERGv13KMkrbP+98IEHCDpK2ApYCXImLnfDwDywsdEf8raSxwAPBInunvdxFxT5XjXEnSo8C/gJ9ExPgKaZaJiJfz438Cy+THywPPF9K9kJe9TAW9HeyioxqtvPXm+rBmV+tnoFk+K3UNIBQRj0paWtJnSCe/tyLieUmDC8nuA34saQXguhpC2HYk/XP56hZgE2AoMCFH4Js351Xu65K+Qzr+5fI2UzpxjD8AtgcezUkXJjUKxgP/J+lU4MYqJ2wiYhpwtKT/Bb4B3CRpbEQcWpb0ZWDFiHhD0gbAnyStWYpaWCXvkNSpWQAbaQINT/jRmuvDrLbvMH9Wku76ieA1wO7AsqSr5lYi4veSHgB2Bv4q6cCIuLNaZh1M/07hsYDbIuIb1fKWtBJwJDAsIt6SNIY03397Kh2jgF9ExPkV9rM+sBNwkqQ7IqK8NwKllsoIUm/ARsDZwEXl6XLcgffz44clPUMKTfxQWdJXJC0XES/n7v5X8/IXgc8W0q2Ql5mZWRPrroGBVwN7kk6S15SvlPQ54NmIOBv4M7BOWZJ/A4t0IH019wObS1ol57OQUhS/okVJDYc5kpYhDZqrWI4ylY7xFuCA0rgDScsXegzejYjLgdOAT43GVwoL/BRwMPB7YI2I+GlEPFch7VL51kWpboYAz1Yo4w3AfvnxfqS6Ky3fNw+e3ASYU7htYGZmTapbegIiYqqkRYAXq5xcvg7sI+m/pHvVJ5etnwJ8pBQ1bwwwXzvpq5XjNUmjgCslzZcX/4RC2NyImJzvrT9Fuk8+oZDFBcDNkl6KiBGF5RWPMSJulbQGcF++/fA2sDewCnCapI+B/wLfrVDc54AtIuK1Gg5tK+CEXB8fAwdFxJsAki4CzouIh0gDLP8gaXTO/+t5+7+SeiVmAO8C+9ewTzMz6+ccRdAARxHsL1wfrbk+WnN9tGjWuiiPIugZA83MzJpUQ8QOkLQEafKbcttGxBs9XR4zM7P+oCEaAflEv15vl8Ma38BBvT4LdF21OS92M5o4sbdLYNan+XaAmZlZk3IjoAFJGi5ps/ZTfpJ+vTyV8tQ8vXGlaYzNzKzJNMTtgP4sTxikiPi4A5sNJ/0c8d4a078L7BsR0/McBg9LuqVewZHMzKwxuCegF1QIfLRPpaBHOYDR8Xn5Y5JWz9MxHwQcrhRgacv29hcRT5emXo6Il0gzCS5Vr+MzM7PG4J6A3jOENKvfDOA6qgc9ej0i1pf0PVKQpW9JOg94OyJOh09mHzyqwj5mlIdzlrQRKabCM3U5KjMzaxhuBPSe53Lkwy/RdtCj6/L/h4GvVcooIq4ArmhvhzmewGWkqItVbz80WmStjpTXo+ebT6O9n+vN9dGiWeqirlEErdNKgY/aC3r0fv7/EVVer1p6AiQtCtxECqN8f4W0n2ikWbSaddYvq53fHy38eWnhukjcCOh99wPnSFolImZIWghYPiKebmObf5MCIQHt9wRImhe4Hrg0Iq7tpnKbmVmD88DAXpYDCI0iBT2aQroVsHo7m/0F+GqtAwNJgYS2AkblbSZJ8uRLZmZNzj0BvSAiZgFrFZ7fCQyrkG5w4fFDpJ8GknsJag2vTA5pfHlny2tmZv2TGwHWVObM7t9TI/g+Z5kmGfhl1lm+HWBmZtak3AgwMzNrUr4dYJb1hwiDngehjKMImrXJPQF9jKQxknZvP2WH871Z0mxJN3Z33mZm1pjcCGhwkmrtzTkN2KeeZTEzs8biRkAXSPppDgR0j6QrJR0paeV81f2wpPGSVs9px0g6W9K9kp4tXe0r+U3O53Zg6UL+G0j6e87rljztL5LGSTpT0kPAD2opa0TcQZpkyMzMDPCYgE6TNAzYDVgXmAd4hDS//wXAQTls78bAucA2ebPlgC1IkwHdAFwLfBVYjRQ/YBngCeBiSfMAvwZ2iYjXJI0Efg4ckPOaNyI2zGWpOYCQmZlZiRsBnbc58OeIeA94T9JfgPmBzYBrcjAggPkK2/wpB+55QtIyedlWwJUR8RHwkqQ78/LVSBMK3ZbzGgC8XMjr6tKDWgMI1ar25xZ2AAAgAElEQVTRgmp0V3k9qK5/arT3c725Plo0S104gFDPmQuYHRHVpuR9v/BYVdIU10+NiE2rrH/nk4Td3BPQSJPNeHIca4/fHy38eWnhukg8JqDzJgBfljS/pIWBLwHvAjMl7QGf3O9ft5187gZGShqQ7/mPyMunAUtJ2jTnNY+kNStlEBFXRMR6Ff58K8DMzKpyI6CTImIi6b7+FOBvwGPAHGAvYLSkycBUYJd2sroemE4aC3ApKYAQEfEBsDtwas5rEulWQ6dIGg9cA2wr6QVJX+xsXmZm1j/4dkDXnB4Rx0lakHRF/3BEzAR2KE8YEaPKni+c/wdwSKXMI2ISacxA+fLhHS1oRNQSbdDMzJqIGwFdc4GkoaQBgWMj4pHeLpB1Xn8ILuT7nGWaZOCXWWe5EdAFEfHN3i6DmZlZZ3lMgJmZWZNyT4CZ9VvDbhrW20WwbjT7sMa/ZdfXuCfAzMysSbkRUCd5fv/StL5/lTQo/32vkOYzkq7tZP41RxuUtJekKZIey7EL2pu7wMzMmoAbAT0gInaKiNnAIOB7heUv9dCEPjOBrSNibeBEUnwDMzNrcm4E1EjSYElPSbpC0pOSrpW0oKRtJT2ar7IvljRfhW1nSVoSOAVYWdIkSaflPB/PaQZIOl3S4/mq/ft5+bGSJublF6gQlKBWEXFvRLyVn94PrND5mjAzs/7CAwM7ZjVgdERMkHQxcARwILBtRDwt6VLgu8CZVbY/BlirFFtA0uDCuu8Ag4H1IuJDSYvn5b+JiBNy+stI0xP/pZippDNomW646KqIOKVs2WjSDIdVNVpQjUYrb725Pqy/6u73drN8VhxAqPs8HxET8uPLgZ8CMyPi6bxsLHAw1RsBbdkOOC8iPgSIiDfz8hGSfggsCCxOmoq4VSMgIg6vZQeSRpAaAVu0la6RJpvx5DituT6sP+vO97Y/K4kbAR0TZc9nA0vUa2eS5gfOBTaMiOclHUeanbA8Xbs9AZLWAS4CdoyIN+pVZjMzaxxuBHTMipI2jYj7gG8CDwEHSlolImYA+wB/b2P7fwOLVFl3W87rrsLtgI/zutdzpMLdgU/9mqC9ngBJKwLXAfsUei3MzKzJeWBgx0wDDpb0JLAYcAawP3CNpMdIJ+3zqm2cr8An5EF+p5Wtvgj4BzAlRw38Zv5FwYXA48AtwMROlvtYUo/FuXlQ4kOdzMfMzPoRpSB21p48iO/GiFirl4tSF3PmzGnIN4Lv67Xm+mht0JmDersI1o26c8bAZv2sDBw4sNUvzHw7wMz6rYk7T2zKL/pqmvXEZ9W5EVCjiJgF9MteADMza05uBJhZv+UAQn2TAwH1HR4YaGZm1qTcCDAzM2tS3dIIkDRK0m+6I69CnrtKGlp4foKk7bpzH2X7K83vj6R7C8tPkzQ1/19K0gM5VsCW9SpLPUjaT9L0/Ldfb5fHzMx6X18eE7ArcCPwBEBEHNtTO46IzQpPvwMsHhEfSdoTeCwivlVrXpIGRMRH3V7IDsgTD/0M2JA06+HDkm4oBBUyM7MmVFNPgKS9JT2YJ5o5P0e821/S05IeBDYvpG0V517S24XHR+doe5Mllaaz/XaOkjdZ0h9zZL7NgK8Ap+V9rlzMt1rkvnw1f7ykR/K61ds4piUk3Zqv8i8CVFj3dv5/A7Aw6aR5NPBLYJdcpgUkbS/pvry/a/KsfqVynCrpEWCPXP6bJT0saXypXPmYzpZ0r6Rny+qtUl1VzKcGXwRui4g384n/NmCHGrc1M7N+qt2eAElrACOBzSPiv5LOBfYGjgc2AOYAdwGPtpPPjsAuwMYR8W4hSt51EXFhTnMSKUrfr/MJ+MaIuDavK+UzPzCG6pH7Xo+I9SV9DzgSqHbV/jPgnog4QdLOpMA6rUTEVyS9XYj69wppHv9D8q2DnwDbRcQ7uZFwBHBC3vyNiFg/b3cHcFBETJe0MSkewDY53XKkgD6rAzcA17ZRVxdUykfSXsBRFY5xRkTsDiwPPF9Y/kJeVlGjRdZqtPLWm+vD+rq+8h7tK+Wot65GEdyWdLKfmE/ECwCbAeMi4jUASVcDq7aTz3bAJRHxLrSKkrdWPvkPIl1139JOPqvRduS+6/L/h4GvtZHPVqX1EXGTpI52jW8CDCVNAwwwL3BfYf3VALl3YDPS1MKldfMV0v0pIj4GnpC0TF72qbpqK5+IuAK4ooPlr6qRJhPx5CetuT6sEfSF96g/K0ktjQABYyPiR58skHal+gn2Q/JtBklzkU6ObRkD7BoRkyWNAobXUKa2vJ//f0R9xzyI1MX+jSrr38n/5wJml3oTKni/8FhV0rSZTw09AS/Sul5XAMa1sS8zM2sCtYwJuAPYXdLS8Mkgs0eBrfN99XmAPQrpZ5F6DiDd158nP74N2F/SgoV8IEXVeznns1chn2oR96YBgyWtkp+3F7mvmrtJkQBLtyoW6+D29wObl8ohaSFJn+oNiYh/ATMl7ZHTSdK67eT9qbpqK5+IuCIi1qvwVxpjcAuwvaTFJC0GbE/7PS5mZtbPtdsIiIgnSPe+b5U0hXSCWg44jtT9PQF4srDJhaQGwmRgU/IVcUTcTLrn/ZCkSaT79QA/BR7I+TxVyOcq4Kg8AHDlQnneowOR+9pwPLCVpKmkXo1/dGTjfCtkFHBlrpf7SPf1K9kLGJ3rZCrpfn9beVerqw7lU8jvTeBEUhTCicAJhdsxZmbWpBxF0ABHEewvXB+tuT5ac320aNa6KI8i6BkDzczMmlRfniyoW0jaH/hB2eIJEXFwb5THzMysr+j3jYCIuAS4pLfLYWY9z1EE+yZHEew7fDvAzMysSTV8I0B1DvxTzL83SVo9T1H8vqQj20i3Uj7WGZKuljRvXj5ffj4jrx/cU2U3M7O+qeEbAUUVAv+sExFHkWY9fCwiPh8R43undEmeT6C9CZQqeRM4FDi9nXSnAmdExCrAW7RMhzwaeCsvPyOnMzOzJlaXRoCkP+UgN1MlfaewfLRy0CFJFyqHH85X6n9UCiQ0UdLmbeTd5cA/VfL9raSHcr7Hl63+oVIwnwcLkwPtIelxpQA/d9dQJxtJOp/0+/6OTkxERLwaEROB/7axD5FiElybF40lRWOENKfA2Pz4WmBbFeYfNjOz5lOvgYEH5PnuFyDFHPgjaZ77nwLrk2YDvBOYnNOfRbp6vUfSiqTZ7NaokneXAv+0UeYf5zIPAO6QtE5ETMnr5kTE2pL2JcUo+BJwLPDFiHhR0qBKGeZZEfcmTW70KnAxcGhEvJ/XX02KhVDuVxFxaRtlrWYJ0tTCH+bnxUBBnwQRiogPJc3J6V8vz6TRgmo0WnnrzfVhfV1feY/2lXLUW1cDCHXGoZK+mh9/FhgCLAv8vTRTnaRraAk6tB0wtHBhuqikhSPibT6tq4F/qvl67rWYmzQj4lCg1Ai4svD/jPx4AjBG0h9oCVr0CUmfAZ4F/gZ8JSKeL08TESO7qezdqpEm0GjWCT+qcX1YI+gL71F/VpJubwRIGk46qW+aw+COA+ZvZ7O5gE3ylMA9TtJKpKl5h0XEW5LG0LrMUf44Ig5SCue7M+m2wwYR8UYh3Suk2ASjgRuUQh5fERGvFvbb3T0BbwCDJM2dewNWIAUPIv//LPCCpLmBgTm9mZk1qXqMCRhIGoD2rqTVSSF3Ic1Zv3UOYjM3sFthm1uB75eeSKoWcQ+6HvinkkVJMQ7mKIXz3bFs/cjC//vyvleOiAci4ljgNdIJ9hMR8VFEXBcRO5NuHywE3J3HSwzMaUZWCfzTmQYAkeaAvgsoBQ7aD/hzfnxDfk5ef2d4zmgzs6ZWj9sBNwMHSXqSFPHvfoB87/xk4EHSSPengDl5m0OBc3IgnrlJJ/qDquR/PCloz1TgXjoY+KeSHMb40Vym50ld/UWL5bK9D5RCB58maQhpYOIdtIxvqJT/i8BJkn4OjOhMGSUtCzxEarB8LOkwYGhE/EvSX4FvRcRLwNHAVZJOIkV7/F3O4nfAZZJmkOp/z86Uw8zM+o8eDSBUus+fewKuBy6OiOt7rABWlQMI9Q+uj9ZcH625Plo0a130dgCh43Jo3MeBmcCfenj/ZmZmlvVo7ICIqDrTXbl6Bf6R9ADp54pF+0TEY13J18zMrNH02QBC9Qr8ExEbd3eeZh0xcFDFaSW6xYZ1y7lBTZzY2yUw69P61bTBZmZmVjs3AszMzJpUUzYCJH1G0rX58XqSdqphm+GSbuzgfgZL+mZ3peusHG/hLklvl+I1mJmZNV0jIM+m91JElCbUWQ9otxHQSYPJExt1U7rOeo8Ut6HmgZlmZtb/9dmBgeUkDSZNRHQ/sBlpBsJLSJMHLQ3slZOeRZry9z/A/hExTdIoUryBhYEBkvYDbiQFMzoBWEDSFsAvSD9d/FQeNZRv67wdpKmFtwJOAdbIP4scS5ob4TLS7IEAh0TEvRXSvUUh4FHugTgdGE+a9GfDvI+LI6IUy6CqiHgHuKcUAbE9jRZUo9HK68F7PavR3h/15vpo0Sx10RsBhOplFWAP4ABSI+CbwBbAV4D/BfYFtsxR8rYDTqZleuL1gXVypMDBABHxgaRjaX3CXbSNPNpyJHBwREyQtDDp6vsY4MiI+FLOe0HgCxHxXp5t8ErSOaE83agq+1gPWD4i1srpBuX/R9HSCCq6OyIOraHsrTTSBBrNOuGH1c7vjxb+vLRwXSSN1giYWfo9f542+I6ICEmPkbrUBwJj8wk2gHkK295WimDYjrbyaMsE4FeSrgCui4gXClERS+YBfpNjI3xESxTFWj0LfE7Sr4GbSDEXiIjTgNM6mJeZmTW5RhsT8H7h8ceF5x+TGjQnAnflK+Uv0zoS4Ds17qOtPKqKiFOAbwELABNy8KRyh5OiC65L6gGYt0p2H9L6tZk/7+OtvO04UmyFiyD1BEiaVOHv7FrKbmZmzanRegLaM5CW0Lmjatzm38AiXcyjFFXwMeAxScOA1UnBiMrzfiEiPs7jEgZUKcMs4HuS5gKWBzbK+1gS+CAi/ihpGnA5uCfAzMw6p9F6AtrzS+AXOSJgrQ2cu4Ch+cp5ZCfzADhM0uM52uB/gb8BU4CPJE2WdDhwLrCfpMmkRkKpd6I83QTSAMUngLOBR3K65YFxeQDh5cCPai2cpFnAr4BRkl6QNLQDx2ZmZv1Qj0YRtL7LUQT7B9dHa66P1lwfLZq1Lno7iqCZmZn1Ef1tTEDd1Su6oZl1v2E3DevtIjSM2YfN7u0iWC9wI6CD6hXd0MzMrKf5dkANJI3q7jn3Je1aHJwn6YQ8OVFdSNpB0jRJMyQdU6/9mJlZ43AjoPfsCnzSCIiIYyPi9nrsSNIA4Bxgx7zPb/jXAWZm5kYAIGlvSQ/mnwmeL2mApP0lPS3pQWDzQtoxknYvPH+78PhoSY/ln/qdkpd9W9LEvOyPkhaUtBlpquPT8j5XLuYraVtJj+a8LpY0X14+S9Lxkh7J6ypNSFTJRsCMiHg2Ij4ArgJ26WK1mZlZg2v6RoCkNYCRwOYRUZrOd29SYKLNSbEJ2r1qlrQj6cS6cUSsS5pvANIUwsPysieB0Tlo0A3AURGxXkQ8U8hnfmAMMDIi1iaN2/huYVevR8T6wG/JUQEljagyY+C9eZvlSRMXlbyQl5mZWRPzwEDYFtgAmJjn+l+AFKVwXES8BiDpatqf53874JKIeBegEKdgLUknAYNIUQxvaSef1UgxEp7Oz8cCBwNn5ufX5f8PkyIjEhF3kYILdYtGi6zVaOWtN9eHdUYzvm+a5Zj7UxTBehAwNiI+mX1P0q7kE2wFn8zrn6f1rTb/f8kYYNeImJyjAw7vYnlL8RI+Ir9+kkYAlUIKvxsRm5GmQf5sYfkKtEyN/CmNNIFGs074UY3rwzqr2d43/qwkTX87ALgD2F3S0gCSFgceBbaWtISkeUjhi0tmkXoOIN3XL0UZvA3YP4cLLuUDKSbAyzmfYrjf8ngBJdOAwZJWyc/3Af7e1gFExF35tkL532Y5yURgiKSVJM0L7Em6HWFmZk2s6RsBEfEE8BPg1jzv/23AcsBxwH2kefyfLGxyIamBMBnYlDz/f0TcTDqxPpTn9j8yp/8p8EDO56lCPlcBR+UBgCsXyvMesD9wTQ6R/DFwXheP8UPgENKtiCeBP0TE1K7kaWZmjc+xAwxw7ID+wvXR2qAzB/V2ERpGs80Y2KyflfLYAR4TYGb91sSdJzblF301zXris+qa/naAmZlZs3IjwMzMrEn5doCZ9VuOIti2ZhsHYJ/mnoBuJGmwpMd7uxzlJM0n6eocPOgBSYN7u0xmZtb73AhoDqOBtyJiFdKkQqf2cnnMzKwPcCOg+w2QdKGkqZJulbSApPUk3S9piqTrJS0GIGmcpA3z4yUlzcqP1ywENJoiaUhe/qlARzWWaRfS9MMA1wLbKs+RbGZmzcuNgO43BDgnItYEZgO7AZcCR0fEOsBjwM/ayeMg4Kwc0GhD4IUqgY72ghTboEoAoX1zfp8EEMoTB80BlujGYzYzswbkgYHdb2ZETMqPHwZWBgZFRGnq37HANe3kcR/wY0krkKIQTpdUKdDRqwARMbI7D6DRgmo0WnnrzfVhtWr290qzHL8DCPWs9wuPPyJFD6zmk2BEwPylhRHxe0kPADsDf5V0IBUCHZXkKIerVcj/VxFxKS0BhF6QNDcwEHijWqEaaTIRT37SmuvDOqKZ3yv+rCRuBNTfHOAtSVtGxHhaBwSaRbq6fxDYvbSBpM8Bz0bE2ZJWBNYBbgX+LOmMiHg1ByhaJCKeq6En4AZgP1IPw+7AneH5os3Mmp4bAT1jP+C8HGHwWVKAIIDTgT9I+g5wUyH914F9JP0X+CdwckS8KakU6Ggu4L/AwcBzNez/d8BlkmYAb5KiCJqZWZNzI6AbRcQsYK3C89MLqzepkP4p0lV+yU/y8lOAUyqkvxq4uhPleo/W4ZDNzMzcCDCz/ssBhFrzfXAr558ImpmZNSk3AszMzJqUbweYWb/lAEItHCzIKnFPgJmZWZNyI8DMzKxJuRFQgaQxknZvP2Wn8z9O0pEVltclFLGkPXJAo49LAYvMzMzcCOgGeSrevrzfx4GvAXfXsThmZtZgGn5goKSfAnsDr5Ei5T0MXA+cAywFvAt8OyKekjQG+BcpMt+ywA8j4tocVvfXwBdyHh8U8t8A+BWwMPA6MCoiXpY0DpgEbAFcCfxfhbINBi4Glszl2z8i/lGWZoOcBtLUwKXlA0gTBg0H5iNFJjxf0nDgROAtYHVg1fbqKCKezHm2lxRovKAajVbeenN9WCWl94XfHy2apS76bQAhScNIoXrXBeYBHiE1Ai4ADsrR9zYGzgW2yZstRzpxr06aU/9a4KukADxDgWWAJ4CLJc1DahzsEhGvSRoJ/Bw4IOc1b0S01b3+a1LQn7GSDgDOBnYtS3MJcEhE3C3ptMLy0cCciBgmaT5ggqRSI2F9YK2ImJnrYTywSIX9HxkRt7dRvooaaTIRT37SmuvDqhkyZIjfHwWui6ShGwHA5sCf87S470n6Cyka32bANYUr3/kK2/wpIj4GnpC0TF62FXBlRHwEvCTpzrx8NdI0wLflvAYALxfyam8K301J3fAAlwG/LK6UNIgUZvjuQpod8+PtgXUKYxMGAkNIvRQPlhoAABGxZTvlMDMz+5RGbwRUMhcwOyLWq7K+GOq3vf5xAVMjYtMq69/paOE6QMD3I+KWVgvT7YB3ypZ1a0+AmZk1h0YfGDgB+LKk+SUtDHyJNAZgpqQ9AJSs204+dwMjJQ2QtBwwIi+fBiwladOc1zyS1uxA+e6lJWLfXsD44sqImA3MlrRFIU3JLcB38y0JJK0qaaFKO4mILSNivQp/bgCYmVlVDd0IiIiJpPv6U4C/AY8Bc0gn09GSJgNTgV3ayep6YDppLMClwH05/w+A3YFTc16TSLcaavV9YH9JU4B9gB9USLM/cI6kSbTumbgol+eR/LPB8+lkz42kr0p6gXR74iZJt7S3jZmZ9X+KiN4uQ5dIWjgi3pa0IOmK/jsR8Uhvl6vRzJkzpyHfCB7c05rrozXXR2uujxbNWhcDBw5sdRu8P4wJuEDSUNKAwLFuAJiZmdWm4RsBEfHN3i6DpB8De5QtviYift4b5TGzpBkDCDlQkHVEwzcC+oJ8svcJ38zMGkpDDwzsyySNK83TL+mveU6AttKfIGm7Opbn8Bw/4HFJV0qav177MjOzxuBGQBfknx+2W4cRsVP+OWBbaY6t10/6JC0PHApsGBFrkSY92rPtrczMrL9zI6CDcqS/aZIuJQXm2UfSfZIekXRNnq+gfJtZkpbMj3+at78nX5EfmZd/ErlQ0raSHpX0mKSL87TBpXyOz/t6TNLqHSj63MACOejQgsBLXasJMzNrdG4EdM4QUjyCrUlz/G8XEesDDwFHVNuoLNbBjqRARuVp5gfGACMjYm3Syfu7hSSv5339Fig1IEZImlTh716AiHgROB34B2na4zkRcStmZtbUPDCwc56LiPslfYkUdGhCji0wL3mioSoqxTootxowMyKezs/HAgcDZ+bn1+X/D5PjEkTEXUC1aZKRtBhpwqSVgNmkuAp7R8TlldI3WmStRitvvbk+mlt7r7/fHy2apS76bRTBXlSau1/AbRHxjR7cdyn2wUfk10/SCOCMCmnfjYjNgO1IDYvXcvrrSDMfVmwENNIEGs064Uc1rg9r6/X3+6OF6yLx7YCuuR/YXNIqAJIWkrRqG+krxTooNw0YXMqTNN3w39sqRETcVSV2QGmK438Am0haUKnLYlvgydoP08zM+iM3ArogX1mPAq7M8QHuA6oO1msj1kExzXukeALXSHoM+Bg4r4vlfAC4Fngk73Mu4IKu5GlmZo3PtwM6KCJmAWsVnt8JfGpasogYXng8uLDq9Ig4rhDr4OGcZlQh/R3A5yvkObjw+CFgeHmaNsr9M+BntaY3M7P+z42AnudYB2Y9ZOLOE33f16wNbgT0sL4Q68DMzAw8JsDMzKxpuSfAzPqt7owi6Oh81h+5J6CBFKcf7uB2x0l6sTCT4E71KJ+ZmTUW9wQ0CEkDupjFGRFxercUxszM+gX3BPQASUdJOjQ/PkPSnfnxNpKukPSNHBDocUmnFrZ7W9L/SZoMbFpYvoCkv0n6do8fjJmZ9RtuBPSM8cCW+fGGwMKS5snLngZOBbYhzf8/TNKuOe1CwAMRsW5E3JOXLQz8BbgyIi4EkDS+SgCh7QplOETSlByVcLG6Hq2ZmTUERURvl6Hfyyf8aaST/HXAVOAq4ETSCX2DiNg3px0NrBkRR0j6EJgvIj7K62aRZhj8ZURc0YH9LwO8DkTe53IRcUAxzZw5cz55IzRLUA3r/7pzYODEnSd2W15mPak4V8bAgQNVXOcxAT0gIv4raSZpiuF7SdMGjwBWAWYBG1TZ9L1SA6BgArCDpN9HbsFJGg8sUmH7IyPi9oh4pbRA0oXAjW2Vt5EmV3EQkNZcH/XTH+rV748WrovEtwN6znjgSNJUweOBg4BHgQeBrSUtmQf/fYO2AwYdC7wFnFNaEBFbVgkgdDuApOUK238VeLw7D8zMzBqTGwE9ZzywHHBfvjJ/DxgfES8DxwB3AZOBhyPiz+3k9QNgAUm/rHHfv8wDD0s9EId36gjMzKxf8e2AHpKDAs1TeL5q4fGVwJUVtlm47PngwtP9O7DvfTpSVjMzaw5uBJhZv+UAQmZt8+0AMzOzJuVGgJmZWZNyI8DMzKxJuRFgZmbWpNwIMDMza1JuBJiZmTUpNwLMzMyalBsBZmZmTcpRBA1oHUXQzMz6p/Iogu4JMDMza1JuBJiZmTUp3w4wMzNrUu4JMDMza1JuBFifJ2lxSbdJmp7/L1Yl3X45zXRJ+1VYf4Okx+tf4vrqSn1IWlDSTZKekjRV0ik9W/ruIWkHSdMkzZB0TIX180m6Oq9/QNLgwrof5eXTJH2xJ8tdL52tD0lfkPSwpMfy/216uuz10JX3R16/oqS3JR3ZU2XuLW4EWCM4BrgjIoYAd+TnrUhaHPgZsDGwEfCz4slR0teAt3umuHXX1fo4PSJWBz4PbC5px54pdveQNAA4B9gRGAp8Q9LQsmSjgbciYhXgDODUvO1QYE9gTWAH4NycX8PqSn0ArwNfjoi1gf2Ay3qm1PXTxfoo+RXwt3qXtS9wI8AawS7A2Px4LLBrhTRfBG6LiDcj4i3gNtKXPJIWBo4ATuqBsvaETtdHRLwbEXcBRMQHwCPACj1Q5u60ETAjIp7Nx3AVqU6KinV0LbCtJOXlV0XE+xExE5iR82tkna6PiHg0Il7Ky6cCC0iar0dKXT9deX8gaVdgJqk++j03AqwRLBMRL+fH/wSWqZBmeeD5wvMX8jKAE4H/A96tWwl7VlfrAwBJg4Avk3oTGkm7x1ZMExEfAnOAJWrcttF0pT6KdgMeiYj361TOntLp+sgXDEcDx/dAOfuEuXu7AGYAkm4Hlq2w6sfFJxERkmr+SYuk9YCVI+Lw8vt+fVm96qOQ/9zAlcDZEfFs50pp/YWkNUld4tv3dll62XHAGRHxdu4Y6PfcCLA+ISK2q7ZO0iuSlouIlyUtB7xaIdmLwPDC8xWAccCmwIaSZpHe70tLGhcRw+nD6lgfJRcA0yPizG4obk97Efhs4fkKeVmlNC/kBs9A4I0at200XakPJK0AXA/sGxHP1L+4ddeV+tgY2F3SL4FBwMeS3ouI39S/2L3DtwOsEdxAGrRE/v/nCmluAbaXtFgeALc9cEtE/DYiPhMRg4EtgKf7egOgBp2uDwBJJ5G+9A7rgbLWw0RgiKSVJM1LGuh3Q1maYh3tDtwZaVKUG4A98+jwlYAhwIM9VO566XR95FtCNwHHRMSEHitxfXW6PiJiy/j/7dy9CcJAGIfxp7S2sA7O4BJauIPiCHZaiIW1A7iBK+gCVhaC+DWG9VnkihRWgpz4Pj94IdOYRJAAAACzSURBVBy5wHsc4R8SklKV7xdrYPXPAQCAlJJl/XRRv7vcAzdgB7TzeA/YNM4bU3/odQdGb65TAafS/ZRcD+qnogScgWOuSemePliDAXAFHsAsjy2BYT5uAdvc+wHoNubO8rwL0C/dS8n1AObAs7EXjkCndD8l90fjGgtgWrqXb5d/DJQkKShfB0iSFJQhQJKkoAwBkiQFZQiQJCkoQ4AkSUEZAiRJCsoQIElSUIYASZKCegH2WkVbJj0gIAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56879ad6a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ask(25, 26, 1, 0, 8, 0, 1, 0, 1, 0, 16, 1, 2, 2, 8, 5, 1, 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "_Stephen Anthony Rose_"
   ]
  }
 ],
 "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": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}