vertex-based-tagging.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pandas\n",
    "import numpy\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import roc_curve, roc_auc_score\n",
    "\n",
    "from rep.metaml import FoldingClassifier\n",
    "from rep.data import LabeledDataStorage\n",
    "from rep.report import ClassificationReport\n",
    "from rep.report.metrics import RocAuc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from utils import get_N_B_events, get_events_number, get_events_statistics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reading initial data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import root_numpy\n",
    "data = pandas.DataFrame(root_numpy.root2array('datasets/1016_vtx.root'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define label\n",
    "`label` = `signB` * `signVtx` > 0\n",
    "* same sign of B and vtx -> label = 1\n",
    "* opposite sign of B and vtx -> label = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "event_id_column = 'event_id'\n",
    "data[event_id_column] = data.runNum.apply(str) + '_' + (data.evtNum.apply(int)).apply(str)\n",
    "# reconstructing sign of B\n",
    "data['signB'] = data.tagAnswer * (2 * data.iscorrect - 1)\n",
    "# assure sign is +1 or -1\n",
    "data['signVtx'] = (data.signVtx.values > 0) * 2 - 1\n",
    "data['label'] = (data.signVtx.values * data.signB.values > 0) * 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>iscorrect</th>\n",
       "      <th>tagAnswer</th>\n",
       "      <th>signVtx</th>\n",
       "      <th>tagger</th>\n",
       "      <th>evtNum</th>\n",
       "      <th>runNum</th>\n",
       "      <th>nnkrec</th>\n",
       "      <th>mult</th>\n",
       "      <th>ptB</th>\n",
       "      <th>etaB</th>\n",
       "      <th>...</th>\n",
       "      <th>svp</th>\n",
       "      <th>om_muon</th>\n",
       "      <th>om_ele</th>\n",
       "      <th>om_kaon</th>\n",
       "      <th>om_same</th>\n",
       "      <th>om_vtx</th>\n",
       "      <th>N_sig_sw</th>\n",
       "      <th>event_id</th>\n",
       "      <th>signB</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td> 0</td>\n",
       "      <td> 1</td>\n",
       "      <td>-1</td>\n",
       "      <td> 0</td>\n",
       "      <td> 490856995</td>\n",
       "      <td> 115839</td>\n",
       "      <td> 3</td>\n",
       "      <td> 19</td>\n",
       "      <td>-0.382002</td>\n",
       "      <td>-6.307031</td>\n",
       "      <td>...</td>\n",
       "      <td> 5.458805</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 1.034008</td>\n",
       "      <td> 115839_490856995</td>\n",
       "      <td>-1</td>\n",
       "      <td> 1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td> 1</td>\n",
       "      <td> 1</td>\n",
       "      <td>-1</td>\n",
       "      <td> 0</td>\n",
       "      <td> 822264706</td>\n",
       "      <td> 115839</td>\n",
       "      <td> 2</td>\n",
       "      <td> 10</td>\n",
       "      <td> 2.117417</td>\n",
       "      <td>-2.673835</td>\n",
       "      <td>...</td>\n",
       "      <td> 3.035841</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 1.091844</td>\n",
       "      <td> 115839_822264706</td>\n",
       "      <td> 1</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td> 1</td>\n",
       "      <td> 1</td>\n",
       "      <td>-1</td>\n",
       "      <td> 0</td>\n",
       "      <td>  68238381</td>\n",
       "      <td> 115839</td>\n",
       "      <td> 4</td>\n",
       "      <td> 16</td>\n",
       "      <td> 0.788837</td>\n",
       "      <td>-4.426422</td>\n",
       "      <td>...</td>\n",
       "      <td> 4.479315</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td>-0.263734</td>\n",
       "      <td>  115839_68238381</td>\n",
       "      <td> 1</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td> 1</td>\n",
       "      <td> 1</td>\n",
       "      <td>-1</td>\n",
       "      <td> 0</td>\n",
       "      <td>   3537659</td>\n",
       "      <td> 115839</td>\n",
       "      <td> 3</td>\n",
       "      <td> 38</td>\n",
       "      <td> 0.070327</td>\n",
       "      <td>-4.690033</td>\n",
       "      <td>...</td>\n",
       "      <td> 4.092348</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 1.062726</td>\n",
       "      <td>   115839_3537659</td>\n",
       "      <td> 1</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td> 0</td>\n",
       "      <td> 1</td>\n",
       "      <td>-1</td>\n",
       "      <td> 0</td>\n",
       "      <td>   2079583</td>\n",
       "      <td> 115839</td>\n",
       "      <td> 2</td>\n",
       "      <td> 31</td>\n",
       "      <td> 1.172778</td>\n",
       "      <td>-3.726187</td>\n",
       "      <td>...</td>\n",
       "      <td> 2.617929</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0</td>\n",
       "      <td> 0.949457</td>\n",
       "      <td>   115839_2079583</td>\n",
       "      <td>-1</td>\n",
       "      <td> 1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 56 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   iscorrect  tagAnswer  signVtx  tagger     evtNum  runNum  nnkrec  mult  \\\n",
       "0          0          1       -1       0  490856995  115839       3    19   \n",
       "1          1          1       -1       0  822264706  115839       2    10   \n",
       "2          1          1       -1       0   68238381  115839       4    16   \n",
       "3          1          1       -1       0    3537659  115839       3    38   \n",
       "4          0          1       -1       0    2079583  115839       2    31   \n",
       "\n",
       "        ptB      etaB    ...          svp  om_muon  om_ele  om_kaon  om_same  \\\n",
       "0 -0.382002 -6.307031    ...     5.458805        0       0        0        0   \n",
       "1  2.117417 -2.673835    ...     3.035841        0       0        0        0   \n",
       "2  0.788837 -4.426422    ...     4.479315        0       0        0        0   \n",
       "3  0.070327 -4.690033    ...     4.092348        0       0        0        0   \n",
       "4  1.172778 -3.726187    ...     2.617929        0       0        0        0   \n",
       "\n",
       "   om_vtx  N_sig_sw          event_id  signB  label  \n",
       "0       0  1.034008  115839_490856995     -1      1  \n",
       "1       0  1.091844  115839_822264706      1      0  \n",
       "2       0 -0.263734   115839_68238381      1      0  \n",
       "3       0  1.062726    115839_3537659      1      0  \n",
       "4       0  0.949457    115839_2079583     -1      1  \n",
       "\n",
       "[5 rows x 56 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Events': 427102, 'tracks': 427103}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_events_statistics(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.397426963471 0.000731430257878\n"
     ]
    }
   ],
   "source": [
    "N_pass = get_events_number(data)\n",
    "tagging_efficiency = 1. * N_pass / get_N_B_events()\n",
    "tagging_efficiency_delta = sqrt(N_pass) / get_N_B_events()\n",
    "print tagging_efficiency, tagging_efficiency_delta"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define B-like events for training and others for prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Events': 247173, 'tracks': 247173}"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sweight_threshold = 1.\n",
    "data_sw_passed = data[data.N_sig_sw > sweight_threshold]\n",
    "data_sw_not_passed = data[data.N_sig_sw <= sweight_threshold]\n",
    "get_events_statistics(data_sw_passed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "features =  ['mult', 'nnkrec', 'ptB', 'vflag', 'ipsmean', 'ptmean', 'vcharge', \n",
    "             'svm', 'svp', 'BDphiDir', 'svtau', 'docamax']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Find good vtx to define sign B\n",
    "\n",
    "trying to guess sign of B based on sign of vtx. If the guess is good, the vtx will be used on next step to train classifier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-folding random forest selection for right tagged events"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_sw_passed_lds = LabeledDataStorage(data_sw_passed, data_sw_passed.label, data_sw_passed.N_sig_sw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 2min 12s, sys: 2.09 s, total: 2min 14s\n",
      "Wall time: 34.7 s\n"
     ]
    }
   ],
   "source": [
    "base = RandomForestClassifier(n_estimators=300, max_depth=8, min_samples_leaf=50, n_jobs=4)\n",
    "est_choose_RT = FoldingClassifier(base, features=features, random_state=13)\n",
    "%time est_choose_RT.fit_lds(data_sw_passed_lds)\n",
    "pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n"
     ]
    }
   ],
   "source": [
    "report = ClassificationReport({'rf': est_choose_RT}, data_sw_passed_lds)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ROC AUC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('rf', 0.55305902958521591)])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "report.compute_metric(RocAuc())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<rep.plotting.FunctionsPlot at 0xda02910>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEiCAYAAADTSFSPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9397990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot([0, 1], [1, 0], 'k--')\n",
    "report.roc()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5 </th>\n",
       "      <td>   ptmean</td>\n",
       "      <td> 0.428948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4 </th>\n",
       "      <td>  ipsmean</td>\n",
       "      <td> 0.261835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7 </th>\n",
       "      <td>      svm</td>\n",
       "      <td> 0.239577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6 </th>\n",
       "      <td>  vcharge</td>\n",
       "      <td> 0.204398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9 </th>\n",
       "      <td> BDphiDir</td>\n",
       "      <td> 0.160545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>    svtau</td>\n",
       "      <td> 0.146701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8 </th>\n",
       "      <td>      svp</td>\n",
       "      <td> 0.144029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0 </th>\n",
       "      <td>     mult</td>\n",
       "      <td> 0.127766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2 </th>\n",
       "      <td>      ptB</td>\n",
       "      <td> 0.118018</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>  docamax</td>\n",
       "      <td> 0.094645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3 </th>\n",
       "      <td>    vflag</td>\n",
       "      <td> 0.039639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1 </th>\n",
       "      <td>   nnkrec</td>\n",
       "      <td> 0.033898</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     feature  importance\n",
       "5     ptmean    0.428948\n",
       "4    ipsmean    0.261835\n",
       "7        svm    0.239577\n",
       "6    vcharge    0.204398\n",
       "9   BDphiDir    0.160545\n",
       "10     svtau    0.146701\n",
       "8        svp    0.144029\n",
       "0       mult    0.127766\n",
       "2        ptB    0.118018\n",
       "11   docamax    0.094645\n",
       "3      vflag    0.039639\n",
       "1     nnkrec    0.033898"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imp = numpy.sum([est.feature_importances_ for est in est_choose_RT.estimators], axis=0)\n",
    "imp = pandas.DataFrame({'importance': imp, 'feature': est_choose_RT.features})\n",
    "imp.sort('importance', ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distributions of output\n",
    "Using flattening of output with respect to one of classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from utils import plot_flattened_probs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEPCAYAAABWc+9sAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x76eee90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEPCAYAAAC5sYRSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x416c250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs = est_choose_RT.predict_proba(data_sw_passed)\n",
    "flat_ss = plot_flattened_probs(probs, data_sw_passed.label.values, data_sw_passed.N_sig_sw.values, label=1)\n",
    "flat_os = plot_flattened_probs(probs, data_sw_passed.label.values, data_sw_passed.N_sig_sw.values, label=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEACAYAAACnJV25AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUUJGAgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x93c7e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "hist(probs[data_sw_passed.label.values == 1][:, 1], bins=60, normed=True, alpha=0.4)\n",
    "hist(probs[data_sw_passed.label.values == 0][:, 1], bins=60, normed=True, alpha=0.4)\n",
    "pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Select good vtx\n",
    "\n",
    "leaving for training only those events that were not-so poorly predicted by RandomForest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mask = ((flat_ss(probs[:, 1]) < 0.6) & (data_sw_passed.label == 0)) | \\\n",
    "       ((flat_os(probs[:, 1]) > 0.2) & (data_sw_passed.label == 1))\n",
    "data_sw_passed_selected = data_sw_passed[mask]\n",
    "data_sw_passed_not_selected = data_sw_passed[~mask]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Events': 185460, 'tracks': 185460}"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_events_statistics(data_sw_passed_selected)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_sw_passed_selected_lds = LabeledDataStorage(data_sw_passed_selected, \n",
    "                                                 data_sw_passed_selected.label, \n",
    "                                                 data_sw_passed_selected.N_sig_sw)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DT for good vtx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 24.8 s, sys: 2.16 s, total: 26.9 s\n",
      "Wall time: 3min 49s\n"
     ]
    }
   ],
   "source": [
    "from hep_ml.decisiontrain import DecisionTrainClassifier\n",
    "from hep_ml.losses import LogLossFunction\n",
    "\n",
    "tt_base = DecisionTrainClassifier(learning_rate=0.02, n_estimators=9000, depth=6, pretransform_needed=True, \n",
    "                                  max_features=8, loss=LogLossFunction(regularization=100))\n",
    "tt_folding_rf = FoldingClassifier(tt_base, n_folds=2, random_state=11, ipc_profile='ssh-ipy',\n",
    "                                   features=features)\n",
    "%time tt_folding_rf.fit_lds(data_sw_passed_selected_lds)\n",
    "pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Report for selected vtx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n"
     ]
    }
   ],
   "source": [
    "report = ClassificationReport({'tt': tt_folding_rf}, data_sw_passed_selected_lds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Default prediction\n"
     ]
    },
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<rep.plotting.FunctionsPlot at 0xe5960d0>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaIAAAEiCAYAAAC7u9a3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8FPX9x/FXCEQRERJRDhUBbzwIlwdYDVat4q1FsYUW\nsFrQetbqz4pKrUWpFtGK2KrghQpq632glYiKgAeoiBdCvEAORe4jwP7++My6k80mmcDOznd338/H\nYx7ZnZ3dfe8m2c9+j5kBERERERERERERERERERERERERERERERGRrDIM+DDqECIiEq77gGeiDlGD\nJkBx1CFEsk2DqAOI1FPMWzKpUcDtVgPLwgwSoqKoA0j+UiGSbFPgLTXpCDwHrAAWAQ8DLX23dwcm\nAUuA5cDrwKFJj7EZOB/4D7AKGE6i260v8IX3+P8FdvTdL75N3H1Y6+1i4BvgB2As0Ni3TRPgAWAl\nsAC4HHgWGFfLa8TL/KqX70fgf0Br77Zy4J9J28ez4NvmTuAWYDHwBjAeeDzpfg2Ar4FLvOsFwBXA\nXGAN8AHw66T7XAtUAOuAhcD9dbwWyXMqRJJLWgNTsA/H7sDPge2Bp0gUr+2xD8bDvW1mAc8DJUmP\ndR1WEA4ARmOtsHZAH+AU4FigM/C3OjL9DCuOPwfOAk7DClPcP4AjgFOBo4GuXrbaWn2dgMnAZ0AP\n4BCs4Db0bk/Vaky1rp+37nCgP/AQcAKwg2+bI4FWwCPe9RuAgVih3g+4EfgX0Nu7/Qzgj8AQYE/g\nRGB6La9FRCTr3EfNY0TXA68krSvGWjjda7hPAdYS8X+r3wzclrTdMGAt0NS37s/A50nbJLeIvqRq\nC+7fwMve5e2B9cCZvtu3I9Fyqsl44M1abp8M3J607j6qt4hmJW3TEPgOGORbdw/wone5CdYK6pl0\nv1FYKxTgMuATEkVRpE5qEUku6Yq1Llb6lq+wb/0dvG12xr7Bf4p1aa3w1u2W9FjvpHj8L73HjFvo\n3bc2c6jaEvHfZw9s/GmG7/Y1wOw6HrMU65bbGjHg3aR1G4EJJIryNsDpWEsJrGW3LfASVd/jwSTe\n34neNvOxIvZLNP4kddC3FsklBVh32uUpblvs/bwf2Akb86gANmDjK8kflqtTPEZl0vUYdX+Z27gF\n96ltDCzINptT3J5qwkWq1/gQ8BbQBhuHKsLGyiCR+0SswPvF35tvgH2wrsijsa7H67DuwzW1ZJY8\npkIk2aim8ZP3sG6ur6heAOJ6AhcCL3jXW5IY5A9DbWM9X2Af4AdjRRGsa+4Aqnb5JZsJHFXL7Uuw\nQuLXCZhXW1DP29hEhLOx8acnSRSQOVhXYjusa68m67Fxt+eBm7Duvh5U7zYVAVSIJDs1wz5Y/d/6\nl2GTCs7FupdGAEuxLqM+2NjFamyAvz/WHbY98HesVRSW2louq7CxoHjW74Ch3n1qK2A3A9OwLsbR\n2Af/z7Aus6+xbrtRwEnY6/09sCvWXebPVVO28dj7uDs2uSJuJTbL7hbvvq9j7+GhwCbgbmAAUIi9\nv6uwCRobqL2wSp6LaozofOyfYi3WF394Hdv3xv7xVmDf9p4E9vLdfjo2JXext8007J9Qck8M+9Cd\nibWA4svN2PhLT6xr6kVsrOUObBpxvNgMwj4838Vmmt1DojVS1/OmKg6xWrYJMnvtcuwD/Wmsi/B9\n7H9iXS1Z3se6vfbF/tanYS3B+Gsc61vewKap/zdAtriHgL2xMbRJSbddg03KuBx7fydhxSre2loG\nnIPNXvzQu+10bHxNxBnxb0jnYH3Jt2PftJIHi+P29La/Cft22wn7kPF/wxqF7dvQzdvmWqxrpq4C\nJ+KabbCW0aVRBxHJZdOxLgW/z7CdBlP5JVZU/N0IvbBvvcn7fiQ/zy1bmFEkU0qBX2FfuDoDj2Jf\nzJLHeERyVqa75oqALlRv7k/CBjNTeRPraz4X63tuivVDz8D2t6jJDnXcLuKKS7Huxf9hM/qOwPZt\nEpEQtMFaMsldZtdiO8HVpAc2/lOJDYq+g/3D1uQCrF+8pu4+ERFxRDbs0NoBm5wwDhsDKsO6LiaS\netbPGdhMqF9hM4hERER+UoS1as5IWj8aOyxJKiOovgf4LljLKrk775fYFN3TawrQpk2b+GwhLVq0\naNESbJlLiDLdItqAFZVjk9YfA0yt4T4FWNHxi1/35z8TO4rxb0nsCV7NggULiMViTi3XXXdd5BmU\nKXcyuZpLmbI3E3Y4qtBE0TU3EptscA529N7bsKP73uXdfiNV98B+GpvgcA2271AXrJvuKxItpb7Y\nTnhXYvtNtPKW2mbVOaOioiLqCNUoUzAuZgI3cylTMC5mClsUhWgidpyvodhOiT2wHVbj4zmtSBxA\nEaywnIUdev897NAs64DjsB1iwfYcb4AVtQW+JfncKiIi4pioDvEzxltSGZhi3ePUXlR6bXWiCA0Y\nMCDqCNUoUzAuZgI3cylTMC5mCluQo/zmmpjX5ykiIgEUFBRAiPUiG6ZvZ0RJSQkFBQU5s5SUbN3w\nWHl5eXre2DRSpuBczKVMwbiYKWw6+rZn2bJl5FJLyfsGIyLivHz8tErZNVdQUJBzhSiXXo+IREdd\ncyIiktNUiCQlF/uplSk4F3MpUzAuZgqbCpGIiERKY0Qel8dU2rVrx9ixYznqqKOoqKigQ4cObNy4\nkQYNav4e4fLrEZHsojEiSVlUVGREJFeoEDmuf//+fPXVV5x00kk0bdqU/fffH4DmzZvTtGlTpk+f\nHsrzuthPrUzBuZhLmYJxMVPYVIiCKChI31JPDz74IG3btuXZZ59l5cqVzJkzB4Dly5ezcuVKDjnk\nkHS/WhGRjNIYkafWMZV07hy6BV1q7du3595779UYkYhEIuwxIh1ZIQh9oIuIhEZdc1nAf7ieTB26\nx8V+amUKzsVcyhSMi5nCpkKUBVq2bMkXX3wBwE477USDBg1+ui4iku00RuRxeUzl6aef5sILL2TF\nihUMHTqUFStWMGbMGCorK3nppZc4+OCDq93H5dcjIhm2cSOsXAmbNtmYd2UlrFsH69fbzw0b7LZY\nzH4mLQXHHw8h1gsVIk+ufXDn2usRyVqxmH3wr10La9ak/um/vH69bV9ZCatXw4oVVkRWrUosq1fb\nsmaNFYvNm1MvlZWJx9sKBVV+pJ8mK0hK5eXllJWVRR2jCmUKzsVcTmWKxWD9espfeYWyrl1rLhCp\nCkXQn/7LmzYFjlYOlKX79RYUQNOm0LChvfZGjWDbbWGbbexnUREUFkKDBrYUFlZdJk1Kd6IqVIhE\nJPts3GgtheXLa15WrLBWww8/wKJF8P338OOPdlu8mypTCgthu+1sady4+k//5aVLYffdrUhstx3s\nsIMVke23TyxNmtjSuLEVlXgBSV4aNrTHKSraut1QQp4kpa45T651ZeXa65EcVllpH76LFsF339nP\npUutgHz/fdXlhx9g2TIrMFurqMhaA3UVhqA/a7utUaOtzxsh7UckItlhzZpEq2PZMvsZX5Yts2Xx\nYluWLbOxjniBqa+CAmspNGtW+9KkCRQXw847Q4sWdrlZM2thZHlxyCUqRJKSU/35HmUKLi25Kiut\naKxaZcVk4UJYsKDqsmRJosDU0Uopp4axj4ICKxKtWkHLlrbstBOUlMCOOyaW+PXiYuuequXIIkG5\n+PtzMVPYVIhE8kUsZuMjixdbAVm82LrAli61LrGFC22Jr1+2rH5HFdlmm0ShKC6G5s2rLkuXQo8e\nVmjixaSkxIpQYWF4r1ucpzEiT0lJCcuWLYsgTjiKi4v54Ycfoo4hmbBxY9VCkrzEb1u82PYXCaqg\nwLq0mjWzbrDWraFNm8TSurXdXlxsLZgddgh9UFuiEfYYUT7+1aQsRCJOW7UKZs+GuXPhiy9g3jyY\nPx8qKuDbb22fkSC23z7R9RVfSkqsW6x1a1vitxcX26wryXuarJAHXOwTVqZg0pZp9Wr46iv48ksr\nMvPm2fX4snBhzfctKLDiES8krVtTvmEDZYcdllgXH39p3Hjrs26hnP79pZGLmcKmQiSSKWvXwocf\nwsyZ8P771pr55htb6uoWbtQI9tsP9t0X9tgD2reHDh2gXTvYbTebiuxXXg559mEm2UtdcyJhWLYM\nZs2yohNfPvmk5p0oi4qgbVtb2rWDPfe0nRrj69q0UTeZREZdcyIui8VsjCa56FRUVN+2QQPo2BE6\nd7Zl771h111tadFCA/2St1SIHOBin7AypbBxI3z2GXzwwU8Fp3zGDMqWL6++7bbbwkEHJYpO585w\n4IEZG6OJ/L1KQZmCcTFT2FSIRFKJxeDzz2H6dJg2DWbMsFlr69ZV37Z586oFp3Nn2GcfdaWJBJSP\nfQEaI5LqYjGYMwdefhleeQXeesuOa5Zs992hUycoLU0UnbZt1a0mOU1jRCJh+e47Kzrx4rNgQdXb\nW7WCQw5JLF26WOtHRNJKpwp3gIvnqM/JTKtWwQsvwB//aOM3rVtD//7wwANWhFq1gn794P77bbLB\nggXw5JNw1VVw1FEpi5CL7xO4mUuZgnExU9jUIpLc9tFH8PjjdmKvGTNswkFc48Zw5JFwzDG2HHCA\nuthEIpCP/3UaI8p1H30Ejz0GEyfCxx8n1jdoAF27WtH5+c+hZ087UKeI1EpjRCJB1FR8iovhtNPg\nlFOs9dOsWXQZRSQljRE5wMU+4azI9NFHcN11tpPoAQfAX/5iRaikBAYNghdftLN93nsvnHxyKEXI\nxfcJ3MylTMG4mClsahFJdlm+3CYT3HVX1ZZPSYm1fPr0sYkFOvumSNaIaozofOBPQCvgI+AS4I1a\ntu8NXAt0BNYDb3r3/9y7vRUwEugM7AU8CAys4bE0RpSN3n4bxoyBCRPslNSQKD5nngm9eqn4iIQk\n7DGiKLrmzgJGATcApcBU4AVgtxq23xN4EjvTcClwNLAt8Lxvm22AJcCNwHRAlSYXbNgADz8MBx9s\ny7hxVoR69YInnrD9gO65B449VkVIJItFUYguA8YB9wKfAhcBC4EhNWxfiuW8CpgHvA+MAPYASrxt\nvgQuBh4Asu60pC72CUeaaelSGD7cTnXw619ba6ikhPIzz7Rjvb36Kpx+uhPFx8XfHbiZS5mCcTFT\n2DJdiIqALsCkpPWTgB413OdNYBVwLlAINAUGADPIwqIjtXjzTduhdNdd4eqrbYfS/feHu++2c/YM\nGQJ77RV1ShFJs0yPEbUBvgGOoOqY0LXAr4B9a7hfD6x7rhgrnjOB47HuuGTPeOsH1fBYGiNyzbx5\ncMkl8Mwzdr2gAI4/Hi691Pb30U6mIpHSfkTQAStC44CHgR2A64GJwFFswXjQgAEDaNeuHQDNmzen\ntLT0p8Oux5vFup6B62vWUD5kCDzyCGWVldC0KeUnnQQnnkjZ2WdHn0/XdT1Pr8cvV6Q6r1YOKAIq\ngTOS1o8GJtdwnxHAu0nrdgE2k7o77xlgbC0ZYq6ZPHly1BGqCTXThg2x2Lhxsdjuu8didtzrWKx/\n/1hs4cLoMm0hFzPFYm7mUqZgXMxEyBPAMj1GtAErKscmrT8Gmz2XSgFWdPzi17VDbjaJxeCRR+zM\npAMHwpdf2ikVXn/dDjzaqlXUCUUkAlF0vp+J7edzPlZ8BmP7/OwPfI1Nwe6OTdMGOBx4DRgGPIpN\nVhiOjSftB6z1tiv1ft4OLAOuwwrfnKTn9wq8ZNRXX9lkg+e9Wff77GMTEs4+WyeQE3FcLo4RTQR2\nBIYCrYEPsR1Wv/Zub4WNC8W9ge179H/AFcAa4C3gOBJFCOA972cMe8NOAiqSHksybc0auPVWm469\nZo2dSuHmm61FVFgYdToRcUBUXVtjgPbYjqndqTqDbiDVi8fjQDesNdQSOBX4JGmbBt5S6LucFUXI\nP0Doiq3OtHkzPPigtXyGDrUi1KePnQX1d7/boiKUk+9TSFzMpUzBuJgpbOoTkfR7+WW48kqYOdOu\nd+4Mt9xix4ATEUmSjztoaIwoLIsWwcUX2/HgAHbZxbrk+vWzcwGJSFbKxTEiyUUTJsDgwfDjj7Dd\ndnDNNVaUGjeOOpmIOE5fUx3gYp9w4EybN8O110LfvlaEjjsOZs+G//u/tBehrH6fMszFXMoUjIuZ\nwqYWkWy5NWtgwAA7M2qDBjY77sILdUgeEamXfPzE0BhROnz2mc2C++AD2GEH65o77rioU4lICHLx\nfESS7Z55Brp3tyK0117w1lsqQiKyxVSIHOBin3DKTJs3w1/+AiefDCtWwBlnwLvvQseO0WWKmIuZ\nwM1cyhSMi5nCpjEiCWb5cujf31pDBQU2LfvKKzUeJCJbLR8/RTRGVF9z5sBpp9m4UHGxnb5bXXEi\neUP7EUm0Xn0VTjkFVq2Cgw6C//4XOmTFkZNEJEtojMgBLvYJl5eXw5QpcNJJVoT69oWpUyMtQs6+\nTw5yMZcyBeNiprCpEElq//kPHH207Ss0cCCMHw9NmkSdSkRykMaIpLpbb4XLLrPLF18M//iHTtkg\nksfCHiNSIZKq7r3XTtMQvzxoULR5RCRy2qE1DzjTJzxhApx7LgDlF17oXBFy5n3ycTETuJlLmYJx\nMVPYNGtOzOOP2+kaYjG44Qbo2TPqRCKSJ9Q1J3DHHXawUoA//QlGjNCOqiLyE3XNSXhiMfjjHxNF\naPhwFSERyTgVIgdE0iccL0IjR0KjRjYx4aqrfipCLvZTK1NwLuZSpmBczBQ2jRHlo1gM/vpXm6bd\nsKGND518ctSpRCRP5WMfTH6PEa1aZbPhHnvMrk+YAGeeGW0mEXGa9iNKv/wtRLEYnHgiPP88NG0K\n48bZqRxERGqhyQp5IGN9wlddZUWouBhmzKi1CLnYT61MwbmYS5mCcTFT2DRGlA/i+waNGGGH6nng\nAdh336hTiYgA6prLD3feCRdcYDPi/vlPuywiEpDGiNIvvwrRa6/ZUbQ3boT774ff/CbqRCKSZTRG\nlAdC6xOeO9fGgTZutKNp16MIudhPrUzBuZhLmYJxMVPYVIhy1bx5cPzx8P330Ls3/P3vUScSEUlJ\nXXO5aM0a6NYNPv7YTu/9xhs2XVtEZAuoa07q74orrAjtsw+8+qqKkIg4TYXIAWntEx43DkaPtuPH\nPfII7Lhj9JnSRJmCczGXMgXjYqawqRDlkkmTEiezGzECOneONo+ISAAaI8oVkyfbMeOWLoWhQ+2g\npiIiaaD9iNIv9wrRhx9Cjx52QNMjjoCXX4aioqhTiUiO0GSFPLBVfcJLltgpHFatgr59bXJCGoqQ\ni/3UyhSci7mUKRgXM4UtaCE6AOiUYn0noGP64ki9VFbCL38JFRXQvTuMHWvHkhMRySJBm1pTgduA\nCUnrzwYuAA5PZ6iQ5UbXXCwGAwfaYXvatIG337afIiJp5krX3IHA2ynWvw0cVM/nPB+YD6wF3qHu\nItYbmAasAJYATwJ7JW1zJPCu95hfAL+vZ6bsM2qUFaHGje0MqypCIpKlghaiTUBJivXNqV+VPAsY\nBdwAlGItrReA3WrYfk+s8JR72x8NbAs879umvXf9DW+bG4F/AqfXI1ek6t0n/NxzcPnldnn8eDjs\nsOgzZYAyBediLmUKxsVMYQtaiKYAV1P1/EWNvHVT6vF8lwHjgHuBT4GLgIXAkBq2L/UyXgXMA94H\nRgB7kCiMg4FvgIu9x7wHuB+4vB65ssf//mfjQps3w7XXwmmnRZ1IRGSrBG3N7Iu1OFZ6PwuwLrXt\ngSOAOQEeowhYDfQFnvCtvwObDFGW4j6tgY+BK7DitZ23/T7Aod42U7ACdaHvfn2A8UBjrDXnl71j\nRLNmQc+ediy5886Du+6ycwyJiITIlTGiT7CxoIeBHYFi4CFvXZAiBNACKAQWJa1fDLSq4T4LsTGi\nG4B1wI/A/sBJvm1apnjMRVjrrUXAbO5bsgROPdWKUL9+MGaMipCI5IT67Ee0AOuK6w2cAAz11oWp\nAzZGNA7ohrWaVgITyaGdcevsE45P0/7ySzjkELj7bmgQ7i5gLvZTK1NwLuZSpmBczBS2hrXc1gXr\n8trkXa7NewGea6n3WC2T1rfEWj6p/B74GrjSt66ft+4wbLLDd1RvUbUENnrPWc2AAQNo164dAM2b\nN6e0tJSysjIg8UeQyeuzZs2qfftbb6VsyhRo3Zryyy+HadNCzxcXxfuRTddnzZrlVB79/up33cXf\nX52fBxm4Hr9cUVFBJtTWqtiMfcAv9i7XJIZ1uQUxDStu/unVnwGPYa2tZH8HegHdfetaA99iY1Nv\nADcBp2HjRnH/xrrweqbKm1VjRP/6FwweDNtsA1OmwMEHR51IRPJMlMeaawd8hRWhdnU8TkXA5zsT\neBDbl2gqNuNtIFY0vsamXnfHpmmDTYh4DRgGPAo0BYZjkyf2w/YbagfMBu7GClBPYDQ2KeK/KTJk\nTyF6/XU46ig71ff999frVN8iIukS5WSFCqwINcIKR8xbl2oJaiJwCTa+NBPogY05fe3d3gobF4p7\nA9v36BSs++8FbNLCcVgRiufsjbWQZmJTvS8kdRFyUnJ3CgBffQVnnGFF6LLLMl6EUmaKmDIF52Iu\nZQrGxUxhq22MKK4SK0R3puk5x3hLKgNTrHvcW2ozBei6NaGcsmGDTU5YsgSOOcbOLSQikqOCNrX+\nAzwLjA0xS6a43zV38cVw++2w++7w3ntQkuqgFiIimRF211yQFhHAK9j4TSfs+HCrk27/TzpD5bXH\nH7ci1KgRTJyoIiQiOS/ozih3ADthYy/3k+guC9JtJnX4qU947lw45xy7fMstkc6Qc7GfWpmCczGX\nMgXjYqawBW0R6QR6YVu3zk71vWKFTVK48MK67yMikgOC9vn9BjsX0fqk9UXYNOkH0hkqZG6OEQ0e\nbPsMdehg40LNmkWdSEQEiHY/Ij//zq1+Lbx12dRicq8QPf00nHKK7bT61lvQuXPUiUREfuLKQU9r\nsht2IFLZUvPmUT7EOwvG8OHOFCEX+6mVKTgXcylTMC5mCltdY0Qf+i6/hh2/La4Q2J2qJ6mT+vj2\nW+jVCxYsgI4drXtORCTP1NXUGub9vBb4B1WnbW/ATvn9hHc5W7jRNbdhA/ToAe++C4ceCi+8AM2b\nR51KRKSaqPcjGub9rMCO9bYurCB5Z+RIK0Lt28Mzz6gIiUjeCjpGdB9WDftgp2Qo9tbviZ0oT+rj\nrbfsNN8Ad91F+ezZ0eZJwcV+amUKzsVcyhSMi5nCFrQQ7YmdsnsM8DcShWgwoAOh1cfixdCnj53s\n7uKL4dhjo04kIhKpoH1+z2InrxuMzZLrBMzDjnh9H1WPmO266MaINm6EX/wCXn0VevaEyZPtUD4i\nIg6LeoworgdwKHaGVb+vgTZpTZTLrrnGilDLlnYcORUhEZF67UdUlGLdbsDyNGXJbU89BTfdBIWF\nMGECtEnUbxf7hJUpGBczgZu5lCkYFzOFLWghmgRclrSuGXA98FxaE+WiGTMSJ7a76SY48sho84iI\nOCRon98uwGTvcntgFjaBYRE2TpR86B+XZXaMaM4c6NYN1q6F00+30zwUhNbVKiKSdq4caw5gO+wA\np129+70HjCdxyu5skblCFIvZ5ISXX7bZcU88Adtvn5nnFhFJE5eONbc9Nh40G/gIGzMaiJ1GXFJ5\n8UUrQs2awfjxNRYhF/uElSkYFzOBm7mUKRgXM4Ut6Ky5s7HThDcAlgHJTYo70xkqJ2zcCJdfbpev\nuQZatIg2j4iIo4I2tb7Ezsx6PVUPfJqNMtM1d8MNVoA6dLBxom22Cf85RURC4MoY0Y9AF2wn1mwX\nfiF65x07zXcsBk8+aecaEhHJUq6MET0KnBBWiJwSi8EFF9jPyy4LVIRc7BNWpmBczARu5lKmYFzM\nFLagY0SXAE8BP8fOUVSZdPv16QyV1aZOtf2GWraEYcOiTiMi4rygTa2LgFHAUmyfoXjfVoF3+cD0\nRwtNeF1za9ZA797w2mtw5ZW286qISJZzZYxoMXATMDKsIBkUXiE67zy4+27YcUeYPRtatQrneURE\nMsiVMaJCrGtOavLpp3DvvXYsuVdeqVcRcrFPWJmCcTETuJlLmYJxMVPY6nNivF+HmCP7XXMNbN4M\ngwZBaWnUaUREskbQptadWCGaDXxAYrJCfIzoovRHC036u+aefRZOOgm23RY+/xx23TW9jy8iEiFX\nzkfUEZjpXd7Xtz5eiPLXDz9YEQK46CIVIRGRegraNVfmW3r5lvj1/PT559Cpk13eeWcYPnyLHsbF\nPmFlCsbFTOBmLmUKxsVMYavPQU/Fb+VKawl98w20bQtPP20TFUREpF7y8cQ46Rkj+vWv4eGH7WCm\nc+faEbZFRHKQK/sR5ZKtL0QzZ0KXLnYg06lT7bKISI5yZT8i8bv9dvt5/vlpKUIu9gkrUzAuZgI3\ncylTMC5mCpsKUX2tXg2PPmqXzzsv2iwiIjlAXXP1sXkznHCCnXn1wAPhgw/Sm0xExEGu7EckGzfa\nVO05c2xiwj33RJ1IRCQnqGsuiOXL4fDDrQgB/PvfduK7NHGxT1iZgnExE7iZS5mCcTFT2KIoROcD\n84G1wDvA4bVsOwzYXMPSwrfdBcDHwBrgE6B/2tJu2gR9+8L06Xb90kuhT5+0PbyISL7L9BjRWcCD\nwBDgDayADMQOIfR1iu2beEtcAXa22M3YSfrwHuvvwO+A6cAhwN3Ar4BnUzxmsDGiykr43e/ggQcS\n60aOtEIkIpJHcm0/ounALOD3vnWfAY8Dfw5w/92w1lQ/rCABTAXeAv7o2+4WrCD9LMVjBCtETz0F\np56auH7HHXYKcBGRPJNL+xEVAV2ASUnrJwE9Aj7GOcAPwBNJj7s+abt1wMHYeZS2zP/+Zz9/8Qt4\n++1Qi5CLfcLKFIyLmcDNXMoUjIuZwpbJQtQCKwyLktYvBoKcRa4QGIR17VX61r/kre+GVexuWDdd\nQ6qOI1W3ahWsXVt9fSxmp3YAO89Qt24B4omIyJbIpunbxwG7YuM/fn/FCtlUrBB9h53I7wpsLKma\nAQMG0G6XXWD0aJo3aULpQw9RVlYGBQX2bWT+fMrmz4eddqJ83TooL7fbSXxbSff1uLAePxeul5WV\nOZUnrjxV6fbrAAAPTUlEQVQDfx+5cF2/v+z5PIhfrqioIBMyOUZUBKwG+lK1a200NlmhrtNJPAWU\nkHrcB6zF1BJYCAwGbgJSHYnUxoguvRRGjbI1JSV2UruOHe3wPR072vqBA2Hs2LpfmYhIDsulMaIN\nwLvAsUnrj8FaM7VpA/SmemvIbxOwADtRX1/gmRq3rKhIFCGwk9stWACvvJIoQgCnnVZHrPRI/hbk\nAmUKxsVM4GYuZQrGxUxhy3TX3EhsjGcGVnwGY91qd3m33wh0B45Out8gYBUwMcVj7gUcCkwDioHL\nsBZWzfsSff557SkPPdRaQyeeWPt2IiKy1aI41twQbPymNfAhcCm2TxHAOOBIoINv+wLgC+B54A8p\nHm9f4GFgH2wSw6vAlUBN1SYWGzfOCk2/fjBoEOy9N+y0E1x/vZ3w7tZboYEOOiEiArl5rLkx3pLK\nwBTrYlQtTMk+waaFB/ftt/Zzl12gl29o6oYb6vUwIiKy9fLza/8339jPXXaJNofHxT5hZQrGxUzg\nZi5lCsbFTGHLz0LkbxGJiEik8vN8RF26wHvvwbRpcMghUecREXFaLk3fdke8RbTrrtHmEBGRPC1E\nixfbrLiWLaNOArjZJ6xMwbiYCdzMpUzBuJgpbPlZiGIxaNUKGmbTEY5ERHJTfo4RAXTvDjNmRJ1F\nRMR5GiMKS9u2UScQERHyuRDtsUfUCX7iYp+wMgXjYiZwM5cyBeNiprDlbyHqUNvBGkREJFPyd4zo\nueegd++os4iIOE9jRGFpFeSksCIiErb8LUSO7EMEbvYJK1MwLmYCN3MpUzAuZgpb/hainXeOOoGI\niJCvY0SHHw6vvx51DhGRrKAxojCMGBF1AhER8eRnISosjDpBFS72CStTMC5mAjdzKVMwLmYKW34W\nIh1jTkTEGfk5RjRzJpSWRp1DRCQraIwoDI51zYmI5DMVIge42CesTMG4mAnczKVMwbiYKWwqRCIi\nEqn8HCOaO9epo2+LiLhMY0RhUItIRMQZKkQOcLFPWJmCcTETuJlLmYJxMVPYVIhERCRS+TlG9N13\nTh19W0TEZRojCoOOrCAi4oz8LESOdc252CesTMG4mAnczKVMwbiYKWwqRCIiEqn8HCNavRq22y7q\nHCIiWUFjRGFQi0hExBn5WYgcm6zgYp+wMgXjYiZwM5cyBeNiprDlZyFqkJ8vW0TERfk5RhSLRZ1B\nRCRraIxIRERymgqRA1zsE1amYFzMBG7mUqZgXMwUNhUiERGJlMaIRESkVrk4RnQ+MB9YC7wDHF7L\ntsOAzTUsLXzb9QfeB1YDC4EHAR3VVEQkC2S6EJ0FjAJuAEqBqcALwG41bH8z0Mq3tAZeAyYDS71t\njgTuA8YCHYFTgf2A8WG8gDC42CesTMG4mAnczKVMwbiYKWyZLkSXAeOAe4FPgYuwFsyQGrZfDSz2\nLUXAz4C7fdt0B74GbgO+BKYDo4FD0h9fRETSLZNjREVYYekLPOFbfwdwAFAW4DGGYV17uwCV3rpO\nWMvqLOA5YEesNbTMe65kGiMSEamHXBojagEUAouS1i/Gut3qUggMwsZ/Kn3r3wf6AY8A673HAxiw\nFVlFRCRD3DroWu2OA3alarccwKHYGNEw4CWgDTa29C/gt6keaMCAAbRr1w6A5s2bU1paSllZGZDo\nn83k9VmzZnHJJZdE9vyprsfXuZLHn8WVPACjRo2K/O8n1fX4Olfy6PcX/LoLnwfxyxUVFeSaIqwl\nc0bS+tHY5IO6PAW8nmL9BKp29QH0xGbWtUmxfcw1kydPjjpCNcoUjIuZYjE3cylTMC5mAkIdz8j0\nfkTTsK603/vWfQY8Blxdy/3aYBMRzgEeSLptIvY6+vjWHQa8CbQFvkna3ntfRUQkiFwaIwIYiY3d\nnINNsb4NGx+6y7v9RuCVFPcbBKzCik6yJ4FTgMFAB6w1dDvwLtWLkIiIOCbThWgicAkwFJgJ9AB6\nY9OvwYpSh6T7FGCFaDywLsVjPgxcDPwB+NB7jk+w4pQV/P2yrlCmYFzMBG7mUqZgXMwUtigmK4zx\nllQGplgXo3pxqs9jioiIw3SsORERqVWujRGJiIhUoULkABf7hJUpGBczgZu5lCkYFzOFTYVIREQi\npTEiERGplcaIREQkp6kQOcDFPmFlCsbFTOBmLmUKxsVMYVMhEhGRSGmMSEREaqUxIhERyWkqRA5w\nsU9YmYJxMRO4mUuZgnExU9hUiEREJFIaIxIRkVppjEhERHKaCpEDXOwTVqZgXMwEbuZSpmBczBQ2\nFSIHzJo1K+oI1ShTMC5mAjdzKVMwLmYKmwqRA3788ceoI1SjTMG4mAnczKVMwbiYKWwqRCIiEikV\nIgdUVFREHaEaZQrGxUzgZi5lCsbFTGHLx+nbs4BOUYcQEcki7wOlUYcQERERERERERERERERyUbn\nA/OBtcA7wOFpetwjgKeBb4DNwG9TbDMM+BZYA0wGOibdvg3wT2AJsAp4CtglaZti4EHgR295AGhW\nQ6argLeB5cBiL9/+Eee6ABvwXO4tU4HeEeZJ5Srsd/jPiHMN83L4lwURZwJoDdyP/U2tBT7C/v6j\nylVB9fdpM/Csd3tBhvMANASGA/Ow92ge8FegMGm7TOcCaAqMwt63NcCbQDcHcuWNs4ANwDnAPsDt\nwEpgtzQ89vHADcAZwGrgN0m3XwmsAE7DisEE7Be9vW+bMd66nwOdsT+AmVSdXv8C8CFwCHAoMBsr\nMKm8iBXEjsABwH+AhdgfSFS5TgZ+AXQA9sTesw0kZjBG8T75HYp9aMzC/j7iosg1DJgD7Oxbdow4\nU3Ps/bkP+/DaHegF7Bthrh2p+h6VApuA/hHlAbgW+B44AWgLnORdH+rbJqq/9Qkkvjx0AK7DCkWb\niHPljenAv5LWfYZ9c0mnlVQtRAVYAbjKt25b7Jd9nne9GbAeONu3za7YP9Sx3vX9sG96h/m26emt\n2ztAribARuyfw6Vc3wPnOpCnGTAXOBL7x4oXoqhyDcP+kVOJKtNw4PUaMkWZy+9q4AfsW3tUeZ4B\nxiWtu99bT4S5GgOVWGH0ewdrsRFRLiA/dmgtAroAk5LWTwJ6hPzc7YGWSc+9Dpjie+6uQKOkbb4B\nPibxyzwMawa/5dtmKtYC8//Ca7ID9rte5kiuQqAv9oc+xYE8/wYeA16j6r51UebqgH3znAc84mWJ\nMtOpwAzsW/Ii7FvwBb7bo/4dFmA9Hg9hH5ZR5XkBOArreQHrlegFPOddjypXQ+z/bn3S+nVYoYj0\n95cPhagF9gtYlLR+MdAq5OeOP35tz90K+0bxfdI2i5K2WZJ0e4zgr+E27IMj/scRVa4DsT/SddiH\n/5nApxHmAWuRdSDRdeI/WVVUuaZhXau/8PK1wv6ZSyLM1AEbZ52Lffu9DbiJRDGK+m/9GKAdcHfE\nee4ExmMfzhuwbqn7gLsizrUS+/8finXFFQL9sK6z1hHmAqxKSjTqOjtfuo56MRL7RnN4gOckwDZb\nk+sT4CCsid8HeBT7thhVnn2Av2HvzSbf4wV5zDBzvei7PBv7AJmPFafpEWVqgLWIrvauvw/shRWi\n0XXcNxN/6+di+Wrq0vQLM89FwECsxf8RNo5yGzZBYGyEucDGzsZirZhNwLtYa7trxLnyokW0FHvT\nWyatb4n1iYbpO99zJT/3d75tCqk6GJ1qm52Sbi/ABmi/o2a3YhM1jsL+EaLOVYl1Nc0E/ox987+A\nxO8h03kOw1rMH3nZKrGB3POxb7NLI8qVbI2XcU+ie68WYBMo/D7BBuTjjxdFLrzbTibRGooyz9XY\neNpE7Hf2EPZlMD72EuX7NA8ow8aMd8VaQ0XAFxHnyhvTSD1Z4W9pfp5UkxUWUH0AcDn2DQ5qHwA8\nxrueagCwh7durxqy3OY99z4pbosyl9+r2EAuEeVphvXhx5f9sW/VD3nXXXmftsUKULz7MIpM47Hx\nAr+/Yi02iPa9usJ7nu1866LKs5iqY2d4GeZGnCuVYmzc+HfedVdy5awzsTfwHOyNug2bDZKO6dtN\nsGmjpdiA3DXe5fhjX4FNkTwNm0r9KNY0buJ7jDuBr6k6JfI9qjZ5nwc+wL7FHIZ1QTxVQ6bR2B9Q\nL6xfNr74nzPTuW7CusDaYWNFN1L1DziK9ymVcqruRxRFrluwlll7bArss16GKP+mumGtxD9jLbM+\nXoYhvm2iyFWAfalM/qIZVZ5/e4/XG/tbPw0rTjdHnAtsbO947O/qGGxXhakk9nFy5X8wpw3B+tnX\nYTt7pmuH1jISO9Jt8l329wdfh33bWEvqncSKsCnDS7FilmonsebYTmLxHUIfwGbDpZKcJb5cm7Rd\nJnONw7oH12GDm5NIFKEo8tTEP307qlyPYDPm1mMfBI9RdX+dKDKBfbjO8p7zE+APKbbJdK5e2N96\n8o6ZUeVpgn2RmI91qX6B7TNXFHEusC8Pc7H/wQXe4zd1IJeIiIiIiIiIiIiIiIiIiIiIiIiIiIiI\niIiIiIiIiIhkqwHY8QdFpBb5cPRtERERkZx2BHaE95XYQSOnY0dgrulYf/2w4x2uwI67NxE7WZnf\nCdhJA+PH/DrLe4y2vm16YGeVXY0dk+5Oqh87TEREclxD7FD6f8eOarw3dlK0g7CTpK3CzsWyM4mj\nGA8EjsOOztwdOx3Ga77HbIsd8PQW7ND5ZwBfYgf3jBeiA7HCdymwB3AwdiTlx9L+CkVExGklWEvl\niBS3DSDYGNG+3mPEW0U3YidV87uKqi2iB4B7krYp9bZpEeA5RZyhMSKRrfMDcB/wEnbeoEup+zxX\nXbDD51dg3XNve+vjRWZf37q4GUnXu2JdfCt9yxvYaZ33qN9LEImWCpHI1huEncBuCnbK6k+xk5Cl\n0gQrWquwQtIN66aDxDlrYlQ90VgqBdipsTv5loOwrrz3t+RFiIhI7ngeO7X22djJ0fy6Yt1nu/vW\nnU7V7r3h1N019xA2iUFERPJcO+w06IdhxaUXNoPtz966zcDR2LhNY2AnbCbcLUAHbHbcR1QtRG2x\ns2jeDOyDFar52GSFeLffgdhsuTHYKZv3BE4E7grpdYqIiKN2Bp7Ais86bHbbTUChd/udwBKqTt8+\nEztl81ps2vexWJHxT3jwT99+DZtptxkrZHFdgRew0zGvAj4AhqXxtYmIiPzkYmyauIiISEZcgO0b\n1B4ba1oC3BppIhERySsjse6+tcDnwPXYzrMiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiJb4v8B8KY0\nm/49kCIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9397390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "report.learning_curve(RocAuc())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('tt', 0.81554209347063)])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "report.compute_metric(RocAuc())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training on all vtx\n",
    "in this case we don't use preselection with RandomForest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DT full"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 5.93 s, sys: 520 ms, total: 6.45 s\n",
      "Wall time: 51.3 s\n"
     ]
    }
   ],
   "source": [
    "from hep_ml.decisiontrain import DecisionTrainClassifier\n",
    "from hep_ml.losses import LogLossFunction\n",
    "tt_base = DecisionTrainClassifier(learning_rate=0.02, n_estimators=1500, depth=6, pretransform_needed=True, \n",
    "                                  max_features=8, loss=LogLossFunction(regularization=100))\n",
    "tt_folding = FoldingClassifier(tt_base, n_folds=2, random_state=11, ipc_profile='ssh-ipy',\n",
    "                               features=features)\n",
    "%time tt_folding.fit_lds(data_sw_passed_lds)\n",
    "pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Report for all vtx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n"
     ]
    }
   ],
   "source": [
    "report = ClassificationReport({'tt': tt_folding}, data_sw_passed_lds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Default prediction\n"
     ]
    },
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<rep.plotting.FunctionsPlot at 0x1078a6d0>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaoAAAEiCAYAAACobJZDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJztnXl4FFXWh9+wIyiIIsFxJDCyKkNEQVTkC6iouKAgCuNC\nGMcNdMQBdVRG3HHUQXDFcUMQFNw3QAYwigsISBDEEUFREARZlEUUSPr741RPKp3uTgequ6s6v/d5\n6klX1amqX1e66/Q959x7QQghhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgSMW4HF6RYhhBAi\nvYwF3ky3iBjUAfZPtwghgkSVdAsQIgmEnCWVVE/QbjuwOZlCkkiNdAsQlRM5KpGJZDlLLNoAbwNb\ngHXARKCRa38HYDrwI/AzMBvoFHGOYmAg8AqwDbibkrBeX2CFc/5XgQNcx4VtwozFWn/XAKuBTcDT\nQG2XTR1gHLAVWAMMBd4CnonzHnE0z3L0/QTMBBo7+wqAhyLsw1pw2TwK3A+sBz4AJgAvRRxXBVgF\nDHbWs4DrgeXAL8BnwAURx9wCrAR+BdYCz5bzXkQlRo5KVDYaA+9jD88OwIlAXeB1SpxbXezB2dmx\nKQSmAA0izjUccxhHAI9grbgcoA/QE+gOHAncVY6mEzDneSJwPnAO5rjC/AvoApwNnAQc5WiL12ps\nB7wLLAOOA47BHHI1Z3+0Vme0bRc62zoDFwHPAacD+7ls/g/IBp531u8EBmCOvDUwAngc6OHs7w0M\nAa4EDgPOAObGeS9CCJFxjCV2jup2YEbEtv2xFlKHGMdkYS0Zd6ugGBgdYXcrsAPY17XtJuCrCJvI\nFtW3lG4B/hv4j/O6LvAbcJ5r/z6UtLxiMQH4MM7+d4EHI7aNpWyLqjDCphrwA/Bn17YngWnO6zpY\nK+r4iONGYa1YgL8B/6XEaQoRF7WoRGXjKKx1stW1fIe1Gpo5NgdhLYAvsZDZFmfb7yPONT/K+b91\nzhlmrXNsPJZSuiXjPuYPWP7rE9f+X4Al5ZwzFwv77Q0hYEHEtt3AJEqcdk2gF9bSAmsZ1gLeofQ9\nvoKS+zvZsfkGc3LnovyXiIN+0YjKRhYWrhsaZd965++zQEMs57IS2InldyIfptujnGNXxHqI8n8Q\n7t6DY+Ll4BKxKY6yP1pBSLT3+BzwMXAwlgergeXqoET3GdgPADfhe7MaaImFOk/CQpvDsfDkL3E0\ni0qKHJXIVGLlbz7FwmjfUdZBhDkeuBqY6qw3oqQIIRnEyzWtwB7wHTGnCRb6O4LSIcVIFgLd4uz/\nEXM0btoBX8cT6jAPK5Toh+W/XqPEwSzFQpU5WOgwFr9heb8pwD1YOPE4yoZlhZCjEhlLPezB6241\nbMaKHi7Fwlf/BDZgIak+WO5kO1aAcBEWbqsL3Iu1qpJFvJbPNiwXFdb6AzDMOSaeg7sPmIOFMB/B\nHMMJWEhuFRYWHAWcib3fy4FDsHCcW1csbROw+9gEK/4IsxWrErzfOXY2dg87AUXAE0A+UBW7v9uw\nApKdxHe8ohKTrhzVQOwLsQOL83eOY5uDhSkil+4um7wYNi1cNvlR9heh2HgmEsIeyguxFlR4uQ/L\n/xyP/f+nYbmeh7Ey6bAz+jP2cF2AVco9SUlrprzrRnMeoTg2iVTfDcUe+G9gIchF2Pfm1zhaFmFh\ntVaYw5qDtSTD7/Fp1/IBVob/agLawjyHfb9+wkr53fwDKxoZit3f6ZgzC7fWNgOXYNWXi519vbD8\nnhC+IPzr6RIsTv0g9issMlEdJgd7qJyMJZjDizuenufYtIqwcTvifOzXW8MIGyGCRk2sZXVtuoUI\nkanMxcIRbpZhHSajkYM5oaPinDPPsTkgjk0+pauxhAgKucCfsD5HRwIvYJ/lyByTEBlJqkN/NYD2\nlA0VTMcSqfF4BRtF4AOsw2A05mP9XWZgziuS2lgIZxXWXyQ3Ac1C+IFrsfDlTCwq0AX7rAshPOZg\nrOUTmZO6BesAGI0DsC9pR8zJ3YZVa7k7X7YALsN+bXbCksdFEdfphCXI/+hsfxFLnB+2x+9GCCFE\nxrEnjioaD2PJ4ni8jQ2LE4sqzjkiRxcQQgjhI1Jdnr4Ba+k0itjeCKvGSpR5lB7CJRqfYIUbsSjG\nQinN3RsPPvjg0Jo1iqgIIUQFWUGSIlSpzlHtxEp+u0dsPxn4qALnyaX8+Hx5NllYP5tSNmvWrCEU\nCvl+GT58eNo1SKd0Sqc0hhdsuK+kkI4OvyOB8ViL5yNsDLBsYIyzfwQ2OOhJznp/zMEVYq2gM7F+\nWNe7zjkY65e1FCvYuBAbvbqXy2Y4NuzLcmzk578Ch2O5rcCxcuXKdEtICOn0Fun0liDoDILGZJMO\nRzUZK5AYhg1Lsxgb/n+Vsz+bksErwTocDsN6wBdhA4UOwDpihqmOdeY8BOtEvMQ55zSXTT1sVOps\nrHPjp1jlVLSBRYUQQgjfEgoC7777brolJIR0eot0eksQdAZBYygUSuqs2omMwFzZcO65EEKIRMnK\nyoIk+RTNR5UgDRo0ICsrK6OWBg0iJ6z1noKCgqRfwwuk01uk0zuCoDHZaPT0BNm8eTOZ1tJyfgEJ\nIYSv0ZOqLFFDf1lZWRnpqDLtPQkh0oNCf0IIISotclQiqQQlvi6d3iKdQCgEK1bAhAkwdChceSVc\neimMHAnz50NRUfo1BgTlqIQQwkvee8+c0ccfw48/xrZr0QJuuQX69oWqVUvv27QJli+H3bth0SIo\nLoZ162D2bFiyBKpXh3r1oHdv6NMHamT2/K/KUZUlkDmqnJwcnn76abp168bKlStp1qwZu3fvpkqV\n2I1mv78nIQLBqlXwxRfmOF59Fe6911pTAA0bQqdO0KEDHOBMl7dgAUyfDqtX23p2Npxxhjmun3+G\nOXOgoCDhFhfZ2XDaaXDssXatNm3KOr4UkMwclVpUGUI0pyMnJESC/PCDhehmz7aw3M6d1krp1Qvu\nvhvq1i1tX1QEr7wCjz8Os2aVOCaAKlXg5pthwABo2hSiVdfu2gXjx8OIEdZyevLJ0vurVYMjj4Ra\ntex11aqw777mjDp0MJuvvoJHH7UW1jPP2AJmd8YZcNVVZh++/saNpvPAA725ZylELaqyBK5FddFF\nFzFx4kRq1qxJ1apVKS4uZseOHdSpUweAGTNmcMwxx5Q5LhXvqaCggLy8vKRewwuk01tSonP5chg9\nGn791VorO3bAhg2wzz5w2GGweTO8/761Ulq3hiOOgGOOMaczeTIsWEBBrVrkNW4ML71kzikaTZvC\nn/5k5w+FzHG8/josW2b7a9a08+7YYde++244rrx5YB1CIfjsM3jrLQv31asHzZrB6afD/vsD5dzL\nUMhaaB9+aC2xjz+Gb78t2Z+dDa1a2X1ZssScVqdOcPbZcPHFtt8jktmiEmWJOTxIOeOHeLfsATk5\nOaGZM2eGQqFQaOXKlaGsrKxQUVFROZKTP1xUUIZ/kU5vSYrONWtCodGjQ6ERI0KhK68MhapX3+vv\n2rvh11lZodDZZ4dCzz4bCi1bFgr98EMoNGdOKJSbG/v4pk1DoYcfDoU2bfL+vbqo8L385ptQ6MYb\nQ6EDDyytt1atUKhmzZL1hQs91YmGUEopzj0vTbmtDy87z+5BK6dp06Y89dRTylGJ4LNpE7z4ooXD\nCgvhhBOshTR+vLWewmRlQf/+1kLYuNHCZA0bwtatFharXduOPegg+O9/YeFCa3Vs3GgtllNOsZDf\n+vVw1lnQvHlZLTt3wtixZle7toX1fvvN9PTqZa0rv1JcDN99Z++9Th3o2NFCjtOnWw5s9GhPn1vJ\nbFHJUZVlzxxVmpGjEoFi+3Z4+WV7kNaoYQ//9evNmXz8sT1ko3HWWRbGq1bNHEX79qnVLWKiYgpR\nLu7hkPw0NJJyKt4SGJ3vvEPeQQfBfvuVtGjmzrX8yapVMGWKtXyiUb06nHiitZa6doV334XPP4dz\nz/XcMQXhfgZBY7KRo8oQGjVqxIoVK+jWrRsNGzakSpUqrFixgubRwhlCeMmuXRaqe+01C9utX2+O\nJVarKMxxx0GXLtZXqFo1aNTIChe6djUHF+aCC5KrX/ge//z09g+BDP298cYbXH311WzZsoVhw4ax\nZcsWHnvsMXbt2sU777xDx44dyxzj9/ckfEgoZBVkW7ZYHmjmTJg0yVpJbqpUsZzPjh2W38nJsVLp\n1q2t0qxjR3stMgblqFJLIB3VnpCJ70kkkXnz4LLLrMAhklatrN9O8+ZWVt2mjSXwRaUh0walHQh8\ng00ZPx/oHMc2ByiOsnR32eTFsGkR45z9nP1v7qF+UQGCMk6ZdMZh4UJzUJ06mZOqW9dCdJ06wd//\nbh1eP/8cBg2C7t2hQwcK5s1Lvc49IAj/9yBoTDapzlGdD4wCrgQ+AAYBU4E2wKo4x50CLHKtb45i\n0wbY5FrfEMWmGXAvMJsk1vwLEXi2bIGJE61EfMEC21alig2ueuutai2JlJLq0N9coBC43LVtGfAS\ncFMU+xzga6ADsCDGOfOAWUBDYGOca1fHnOPDQDfgQODMKHYK/YnKzezZcM451t8IoH59uOgiuOIK\nC+kJEYVMKU+vAbTHWjRupgPljTfyClAL+Ap4AHg5is18oCawFLgTKIjYfxfm9MYDJ1ZAtxCZSyhk\n4bwpU6yVVFQEN91kHV07dbK8U69e1tlViDSRyhzVgUBVYF3E9vVArAGntgJDgD7AacBMYBLgrldd\nA1wB9HKWLx07d+6rO3AuJS25pA73IUoISny90ugMhayz7Wef2RQTLVta36Rhw+Daay20t3OnOagP\nPrDS8D1wUpXmfqaAIGhMNn7vR7URa0GF+RQ4ALgemOBsW+YsYeZgIcPrsFBfQ2As0BfY4thkEaeJ\nmp+fT05ODgD169cnNzd3r96Enwl/CcIdCr1eL3QqxJJ1fq/Ww/hFjyf3MxSi4O23bb1BA3jwQQpe\new1++4288PsFqF+fvPPPh6pVKSgshGOPJe+++yArS/fTB+uFhYW+0hNeLygoYOzYsQD/e14mi1Tm\nqGoA2zGH4Q7dPYIVQnRN8Dz9gceAfeLYDMcKN9pQksNyT+4SbkkWOzZfufZFzVE1aNCAzZuj1XAE\nl/33359NmzaVbyiCwebNVkK+bJmNaTdjhk22F0mtWjZaRPfucP75kJfn7zHrRCDIlBzVTqwgojul\nHdXJwIsVOE8uFu5L1OYT4AjXviwsh1UfqzpcmchF9UAXviUUgmefhcGDbUoLN3Xq2FxGdevatA5X\nXgmHHpoenULsIanuRzUSyAcuAVoDo7H81Bhn/whghsu+P9bvqTXQEhiK9cN6yGUzGOgJNAcOd87R\nE6vuA/gFK7AIL58DPwPbnPVd3r291BGUuLV0eksZnZs22UCtAwaYkzr6aLj0UhsZe8kSG0/v55/h\n++9tkr4UOanA3k8fEgSNySbV7f3JWI5pGNAYWAz0oKQPVTbW1ylMyLFtgoXuvgQGABNdNtWB+4BD\nsE7ES5xzToujQ8UUIvgsWWIT4K1YYSXkDz4IF17o7ZQzQvgAfaLLEjVHJUTa2LwZnn4aXnjBBmtt\n3hwWL7Y8VHGxTVn+6qvQpEm6lYpKjMb6Sy1yVCL9rFsHY8bYVOoffVR6wsAw1atbR9yHHrIp0IVI\nI5k21p/wgKDEraVzD3jzTWjb1oYqmjXLnNTJJ8Orr1IwYgSMHGmTDm7YAE895Usn5av7GYcg6AyC\nxmSjmlQh/MKKFXDzzTZtBkC3btbx9vjjrZwcbArxSj6Jnqh8KPRXFoX+RGrZtQtuuw3uvdde16xp\nFXrXXGMDwQoRADKlH5UQIpKvvrJhiubNs2q9/v3h9tvV10kIF/q5FlCCEreWzhiEQpZfOvJIc1KH\nHgrvvQdjx8Z1Urqf3hIEnUHQmGzkqIRIJaGQVfKddhr85S82QGy/frBoEZxwQrrVCeFLlKMqi3JU\nwnvWr4fnn4dx4+DTT23bvvvCo49aJ10hAo5yVEIEld27bTijW26BX36xbQ0aWDXfVVdBw4bp1SdE\nAFDoL6AEJW5daXXu2mUjSRx1lM3x9MsvcOqpVnq+erVV+e2Bk6q09zNJBEFnEDQmG7WohPCKNWvg\nySetOGLevJIpNpo0sRBfjx7p1SdEQFGOqizKUYmKsWyZOaLHHy891FHLljZr7sUXayp3kfEoRyWE\n3wiF4PXX4Z//tMFhw/TqZZMRtmtng8eqw64Qe42+RQElKHHrjNQ5e7YNa3TOOeak6ta1+aAKC20M\nvvPOs9ZUEpxURt7PNBIEnUHQmGzUohIiURYvhptugrfesvWGDeEf/4A//9lm0hVCJAXlqMqiHJUo\nzfLlMHy49YMKhcwpDR0KQ4ZYXyghhHJUQqSUUMg66H73nVXxPfUUFBXZ/E+XXw7DhkGjRulWKUSl\nIV05qoHAN9jU8fOBznFsc4DiKEt3l01eDJsWLps+zrU2A9uAhcDFe/tG0kVQ4taB0Tl5ss2i27ev\nhfSys6FjR/j3v81xDRhg1X0PPZRWJxWY+ymdnhEEjckmHS2q84FRwJXAB8AgYCrQBlgV57hTgEWu\n9c1RbNoAm1zrGyJe3w78F9gFnAk8BWwE3q7QOxCZw9KlVj6+YEHp7fvvb4PD5ubC3/8OrVqlR58Q\nIi05qrlAIXC5a9sy4CXgpij2OcDXQAdgQZT9YC2qWUBDzPEkygJgGnCza5tyVJWFN9+0KTa2brXK\nvW7dbCbd7t2ttDxLKVwhEiWTpqKvAbQHpkdsnw4cV86xrwDrsFZY7xg284E1wAzMecUiCzgRCw3O\nLOe6ItP49VcrhDjrLHNSffrADz9Yv6irroIWLeSkhPARqXZUBwJVMYfjZj2QHeOYrcAQLMd0GuZY\nJgEXuGzWAFcAvZzlS8cuMvdVD8tP/QZMAa7FWmKBIyhxa9/pXLUKOnSAkSOhalW4+26YNImCefPS\nrSwhfHc/YyCd3hEEjckmCFV/G4EHXOufAgcA1wMTnG3LnCXMHCxkeB3WAguzBfgjUBc4CRiNOck3\n3BfMz88nJycHgPr165Obm0teXh5Q8qFJ93oYv+iJtV5YWOgfPT//TEGXLrByJXnNm8P48RTs2GET\nFjqk+34F6n7GWQ/jFz1Bvp+FhYW+0hNeLygoYOzYsQD/e14mi1THN2oA24G+wMuu7Y9ghRBdEzxP\nf+AxYJ84NsOxwo02cWyeAA6LuK5yVJnI9u02ksR//mOFER99ZAUTQghPyKQc1U6sgKF7xPaTgY8q\ncJ5cLNy3tzZV0TBSmc0XX9iQRg0bmpM66CCYMkVOSogAkY6H9EggH7gEaI2F37KBMc7+EVgxRJj+\nQD/HtiUwFOuH9ZDLZjDQE2gOHO6coyfwsMvmZqyAoplzriHAhcB4r95YKokMsfiVtOp8803rC/Xi\ni7BjBxx7LEydCk2bljHV/fQW6fSOIGhMNunIUU3GckzDgMbAYqAHJX2osjFnEibk2DYBirBCiQHA\nRJdNdeA+4BCsE/ES55zTXDZ1sHBh2OYL4CKsMENkEps3wx13wKhR1ln3vPPgvvusX5QQInCoBrcs\nylEFmRdegEGDYNMmKzG/80648UaVmwuRZDTWnxCJMHo0DB5sr/Py4F//gvbt0ypJCLH3qJAgoAQl\nbp0Snb/9BtddV+Kk7r8fZs2qkJPS/fQW6fSOIGhMNmpRiWAzd67NB7V0qU1U+MQTti6EyBgUuC+L\nclRB4JdfbNLCUaOguNiGPXr6aZt5VwiRcpSjEsLNqlVwyinWR6pKFbjhBpvYsHbtdCsTQiQB5agC\nSlDi1hXWuXq1zQF1wQXwl79Yq2ncOFi4ENautengO3c2J9W6NcyZA/fcs9dOKmPvZ5qQTu8IgsZk\noxaVSC/FxTaT7nvvwTPPlBp3Ly7HHgtvv60RJoSoBChHVRblqFJBYaGNYP7qq7BtW8n22rVtPqhT\nTrGw3vffW6HEkiXw0082RUf37paPqls3ffqFEKVQjkpkDkVFcMUV8OSTJduys+GPf7R5oc47D/bb\nL336hBC+QzmqgBKUuHUpnUVFMGCAOanata3f04oVlnt65x3LSaXJSQXyfvoY6fSOIGhMNmpRidTw\n00/miF5+GerUgWnTrChCCCHKQTmqsihH5TXvvw8XXmhl5XXr2gjmclJCZBTJzFHJUZVFjspLVqyA\ndu1s4sKOHeG556B583SrEkJ4TCZNnCg8IhBx66IiCs45x5xU797wwQe+dVKBuJ9Ip9cEQWcQNCYb\n5ahEcti92zriLl5sVX2PPw7Vq6dblRAigCj0VxaF/vaGLVtgyBCbVffnn23bm2/CGWekV5cQIqmo\nH5UIBl9/DWeeaR10AVq2hKuvlpMSQuwV6cpRDQS+waaEnw/EKwHLAYqjLN1dNnkxbFq4bC4FZgOb\ngM3ALCCwQ237Km69dSuMGAFHHWVOqnVr+Owz+O9/KTj88HSrSwhf3c84SKe3BEFnEDQmm3S0qM4H\nRgFXAh8Ag4CpQBtgVZzjTgEWudY3R7FpgzmiMBtcr/8PeB74EHOQ1wLvALnA8gq9A2FDG919tw0W\nu3hxyTBIZ5xhlX316qVXnxAiY0hHjmouUAhc7tq2DHgJuCmKfQ7wNdABWBDjnHlYC6khsLECWtYC\ndwKPuLYpR1Ue8+fDWWfZiBJhjj8ebr0VTjwRspT6FKKykUnl6TWA9sD0iO3TgePKOfYVYB3WCusd\nw2Y+sAaYgTmveNQEahG9ZSZiMW0adOliTqpLF5g5E9assdLzk06SkxJCeE6qHdWBQFXM4bhZD2TH\nOGYrMAToA5wGzAQmARe4bNYAVwC9nOVLxy5e7utO59xvVOgd+IS0xK3nzrX+UDt2wCWXwH/+A926\nQePGMQ8JSnxdOr1FOr0jCBqTTRCq/jYCD7jWPwUOAK4HJjjbljlLmDlYyPA6rAUWyTXAZcCJwLbI\nnfn5+eTk5ABQv359cnNzycvLA0o+NOleD5Oy6x98MJx+OgW//AKnnkreE09AVla5xxcWFqZGX9Du\n5x6u6356ux6E+1lYWOgrPeH1goICxo4dC/C/52WySHWcpgawHegLvOza/ghWCNE1wfP0Bx4D9olj\nMxwr3GgTsX0wcDtwKvBRlOOUo4okFIKuXW1Sw9NPh9deg2pB+I0jhEgVmZSj2okVRHSP2H4y0Z1G\nLHKxcF9Fbf6GOakeFbxe5WbKFHNSDRpYRZ+clBAihaSjH9VIIB+4BGgNjMbyU2Oc/SOwYogw/YF+\njm1LYCjWD+shl81goCfQHDjcOUdP4GGXzXXO9kuwcvRsZwnkLH2RIZakUVQEf/+7vb75Zqhfv0KH\np0znXiKd3iKd3hEEjckmHT+NJ2M5pmFAY2Ax1sIJ96HKBpq57EOObROgCCuUGABMdNlUB+4DDsH6\nSC1xzjnNZTMQe7+TIvSMBf68d28pgxk/3qaBb9IEBg1KtxohRCVEtcRlUY4qzK5d0KIFrFwJ48bB\nRRelW5EQwqdkUo5KBInnnjMn1bIl/OlP6VYjhKikyFEFFM/j1s8/D48+atNzgP296y57PWwYVK26\nR6cNSnxdOr1FOr0jCBqTTaI5qiOwjrqLIra3A3YBS70UJVLMHXfALbfY62eftTH83nvPZuc97DDo\n2ze9+oQQlZpE44kfYdV5kYUI/bBBZeONABE0KleO6pZbzFFVqQIHHQQ//FB6/zPPQH5+WqQJIYJD\nMnNUiZ50K9Z6+jpi+2HYSBGBLPGOQeVxVBMnwgUXWFjvuefg1FPhhhvgk0+gWTPo0AGuu26Pw35C\niMqDH4opioAGUbbXR5WDaWGv49ZffAGXXWavH37Ywnv169uU8QsXwssvW/+pvXRSQYmvS6e3SKd3\nBEFjsknUUb0P3EzpnFZ1Z9v7XosSSWLJEpuBt317G/l8+3ZrUV1+efnHCiFEmki0NdQKG9x1q/M3\nC8tL1QW6kFnFFJkX+isuhvvvh3/8A3buLNmemwuzZ0PduunTJoTICPyQowI4GCucOBIbLWIh8Cjl\nj7kXNDLPUY0YATc5c1JedhlceqmN13fEERq3TwjhCX7IUYE5pJuxoYlOx4Y1yjQnFRgSjlt/+KG1\npAAmT7Yc1NFHW2sqBU4qKPF16fQW6fSOIGhMNvGeVO2xflNFzut4fOqZIrH37NoFM2bA8uVw7702\nsOwNN0CfPulWJoQQFSZeM60YGyB2vfM6FiGsM3CmEPzQ3803W6fdMMceax14q1dPnyYhREaTrhxV\nDvAd5qRyyjnPSm/k+IJgO6offrA+UDt2wF/+Aq1a2bTxFZyeQwghKkK6clQrMSdVHZsiI+Rsi7aI\nFBMzbn3PPeakevaEJ56AIUPS6qSCEl+XTm+RTu8IgsZkk0gxxS7MUaljr99ZvRrGOPNP3n57erUI\nIYRHJOp8XgHeAp5Ooha/ENzQ36BBNgL6eefBpMhhGYUQInn4oR/VQGA48AIwH9gesf8VL0WlmWA6\nqnXrICcHfv3VRqA4/PB0KxJCVCL80I/qYaAhcDXwLPBSxFJRBgLfYNPGzyf+6Os5WK4scunussmL\nYdPCZXO4o3WFs2/4Huj2DWXi1g8+aE6qZ09fOamgxNel01uk0zuCoDHZJNrj08sJFs8HRgFXYsMx\nDQKmAm2AVXGOO4XS82FtjmLTBtjkWt/gel0bG/39ZeBOrDgkM9iyBR55xF7fcEN6tQghhMck2ky7\nGJuL6reI7TWAvsC4ClxzLlAIuEdCXYa1dm6KYp+DOZgOwIIY58wDZmGtvo0JaFgMvAhEqzgITuiv\nuBjmzoVRo2zUiS5drL+UEEKkGD+E/sYC9aJs38/Zlyg1sFEupkdsnw4cV86xrwDrsFZY7xg287Fh\nnWZgzitz+eYbOOEEOO44c1JVqpTM0iuEEBnE3ob0fg/8VAH7A7FRLNZFbF+PjYIRja3AEKAPcBow\nE2vdXeCyWQNcAfRyli8du0yaebiE55+n4Igj4KOPbFbewYNtssMTT0y3sjIEJb4und4ind4RBI3J\nprwc1WLX6/eA3a71qkATYIrXoiLYCDzgWv8UOAC4HpjgbFvmLGHmYCHD67AWWIXIz88nJycHgPr1\n65Obm0t/BhwlAAAbdklEQVReXh5Q8qFJy/q2bRScey68844JPeccCvLzYb/9yDvqqPTri7JeWFjo\nKz2x1sP4RY/uZ2rWg3A/CwsLfaUnvF5QUMDYsWMB/ve8TBblxRNvdf7eAvyL0mXpO7HKvZed14lQ\nwzlHX+e4MI9ghRBdEzxPf+AxYJ84NsOxwo02Ufb5L0dVXGx9n7p2hewojcsNGyAvDz7/HGrXhgce\nsCk7stQPWwiRfpKZoyqvRXWr83cl1ofq17283k6sIKI7pR3VyZjjSJRcyp9iJBEb/zB+POTnQ+vW\nViCx774l+7ZsgdNOMyfVqpVNE98mmv8VQojMoyLFFFlYnugGYH9n+2FYGK4ijATygUuA1sBoLD/l\njP3DCKwYIkx/oJ9j2xIYivXDeshlMxjoCTTH+kuNcNYfdtlUx5xXLlaq3th5fVgF9SeHcU7h5Bdf\n2CCy4Vbdhg02ffz8+TbY7KxZ0KZNmRCLX5FOb5FObwmCziBoTDaJ9qM6DHMedYH6WOtnM1bAUB/4\nSwWuORlzbsMwZ7EYm4wx3IcqG2jmsg85tk2wubG+BAYAE1021YH7gEOwTsRLnHNOc9n8jpJ5s0JY\nefzlQAHQrQL6vef77+Hdd6FGDahZE1580ar4unaFO++0MfwaN4b//Mf+CiFEJSLReOJbwFrMMf0E\ntMP6NnXBWlvNYh4ZPFKfo/rXv2DoUOjdGy64AM4913JWYTp1svzVoYemVpcQQiRIOnNUYY4DOmEt\nGjergIM9VRRUfv4ZqlaFunVLb1+2zMJ5Z50Vu/BhglO8eMEFcM458Nln8MYbNo38UUfZRIg1aiRX\nvxBC+JSK9KOK9qT8PfCzR1qCy7Zt0LIldO5cuiUUCpnjOftseDrGwPNffAELF9qcUT162LbDD4cb\nb4S33oLbbovqpIISt5ZOb5FObwmCziBoTDaJOqrpwN8ittXDyrvf9lRREJk710YvX7TI8khhFi2C\npUvt9TXXwFdflT32rrvs77nnWn5KCCFEKRKNJ/4OeNd53RQbq+8wbISJLtjIEplCxXNUd90Fw4bZ\n67PPhldftdc33mgz7taqZSObd+xoY/HVqmX7CwqsYKJWLXNoTZt69iaEECKV+GGsv++xUu57gH8D\n87BRH44ks5zUnvHxxyWv33wT1qyxsF948sLnn4ff/96GOerQARYvhs2bYeBA23/zzXJSQggRg4rk\nqOpi+aglwOdYzmoA1qep8hIKwZw59rpDBygqgqeeggULbODY7GzrB/X229C8uU1q+Mc/QoMGlp9q\n3hyuu67Clw1K3Fo6vUU6vSUIOoOgMdkkWvXXD5uGvgrWfyoyNvaol6ICxYoVsHEjNGpkIcDu3eH+\n++G112x/nz5WDdi2LXz6KVx7rXXurVnTBpR96inlpoQQIg6JxhO/xWb2vZ3SA9NmIhXLUY0fDxdf\nbDPrvvKKVe6FB40F+OADOP74yCtojD4hREbhh35U9bCOvZnupCpOOOx37LE2msTUqTBvns0Rte++\nNl9UJHJSQgiRMInmqF4ATk+mkMASdlSdOtnfrCyr7rv/fhg+PGlOKShxa+n0Fun0liDoDILGZJNo\ni2ow8DpwIjY2366I/dGmy8hcdu2Cxx6zflKLFllL6uij061KCCEykkR/7v8VGAVswMrRw0mcLOd1\nW++lpY3yc1S33Qa33lqyfvzxlosSQohKSjJzVImedD3Wh2pkMkT4jPiOaskSaN/eWlX//KdNvZGX\nBwcemDKBQgjhN/zQ4bcqFvqr3BQV2VxRu3bBFVfA9dfb0EdpcFJBiVtLp7dIp7cEQWcQNCabikyc\neEESdQSDt9+20SUOOcRaU0IIIZJOos20RzFHtQT4jJJiinCO6q/eS0sbsUN/DzwAf/sb/PWvMHp0\nalUJIYSP8UM/qjbAQud1K9f2sKOqHKxda381y64QQqSMREN/ea6lq2sJr1eEgcA32JTx84HOcWxz\ngOIoS/cIbdFsWkScqzewFPgVG6vw7Arqhh9+sL/Z2RU+1GuCEreWTm+RTm8Jgs4gaEw2FRmU1gvO\nx8rc78RGY/8ImIpNwBiPU4Bs1/JuFJs2ETbLXfuOxTotjwfaAROAF4GOFVKvFpUQQqScVI/lMxeb\ny+py17ZlwEvATVHsc4CvgQ7AghjnzANmAQ2BjTFsJgH1MYcX5j/Aj8CfImxj56jatrXy9MJCaNcu\nxqWEEKLy4YfydC+oAbTHZgt2Mx2IMiBeKV7BJmn8AAvhRWM+sAaYgTkvN5328LqlCbeofBD6E0KI\nykIqHdWBWH+sdRHb12OhumhsBYYAfYDTgJlY68hdKr8GuALo5SxfOnbu3Fd2lOuui3PdsuzcadN5\nVK3qi869QYlbS6e3SKe3BEFnEDQmm0Sr/tLFRuAB1/qnwAHA9VieCSx0uMxlMwcLGV6HtcAqTH5+\nPjk5OQDUr1+f3Nxc8v7wBwAK6tWD2bPJy8uzdedDlOr1MOm6fqLrhYWFvtKj+5ma9TB+0RPk+1lY\nWOgrPeH1goICxo4dC/C/52WySGWOqgawHegLvOza/ghWCJFo9WB/4DFgnzg2w7HCjTbO+rfAQ8D9\nLpvrgEGYU3MTPUc1b56Nit6+vc3eK4QQ4n9kSo5qJ1YQ0T1i+8lY9V+i5GLhvorYfOxcJ/K6HyZ8\nVeWnhBAiLaS6PH0kkA9cArQGRmN5ojHO/hFYMUSY/kA/x7YlMBTrh/WQy2Yw0BNoDhzunKMn8LDL\nZjTQDbgB67B8I1ZwMSph5T7qQwXBiVtLp7dIp7cEQWcQNCabVOeoJmM5pmFAY2xuqx7AKmd/NtDM\nZR9ybJsARVihxABgosumOnAfcAjWiXiJc85pLpuPsZDjndjcWcuB84B5CStXHyohhEgLmhO9LNFz\nVFdeCWPGwEMPwVVXpV6VEEL4mEzJUQWbcOhPLSohhEgpclSJ4rNiiqDEraXTW6TTW4KgMwgak40c\nVaL4rJhCCCEqC8pRlaVsjioUglq1bHSKbdugTp30KBNCCJ+iHFW6+eknc1L77isnJYQQKUaOKhF8\nWJoelLi1dHqLdHpLEHQGQWOykaNKBOWnhBAibShHVZayOaoJE+DCC+G882DSpPSoEkIIH6McVbrY\nuRPuuaekg+8hh6RXjxBCVELkqOJx//1w441WTJGXB1dfnW5F/yMocWvp9Bbp9JYg6AyCxmQjRxWP\n1avt7z/+AbNmQZLnXBFCCFEW5ajKUpKj6t8fxo2DZ56B/Py0ihJCCD+jHFW62LbN/tatm14dQghR\niZGjioePHVVQ4tbS6S3S6S1B0BkEjclGjioePnZUQghRWVCOqiwlOap27eCzz2DhQsjNTa8qIYTw\nMcpRpYvt2+2vxvcTQoi0kQ5HNRD4Bps2fj7QOY5tDlAcZekew74zsBub4t5NdeAWbAr6HUAhcEq5\nSn0c+gtK3Fo6vUU6vSUIOoOgMdmk2lGdD4wC7gRygY+AqcDvyznuFCDbtbwbxWZ/YBwwA4icS/5O\n4ArgaqA1MAZ41dEQGx87KiGEqCykOkc1F2vNXO7atgx4Cbgpin0O8DXQAVhQzrlfARZizvdcoK1r\n3xpgBPCQa9tLWOvqoojzWI6quBiqVrUtu3eXvBZCCFGGTMlR1QDaA9Mjtk8Hjivn2FeAdcAHQO8o\n+wcCDbGWU7QbVQP4LWLbr8QLO+7YYX/32UdOSggh0kgqHdWBQFXM4bhZj4XzorEVGAL0AU4DZgKT\ngAtcNm2x/NOFlA35hXkHGAy0wN7zyUCvONctCfv5tJAiKHFr6fQW6fSWIOgMgsZkUy3dAsphI/CA\na/1T4ADgemACUBNzXEOBb+Oc5xrgCWAp5syWA08Df45mnJ+fT85++wFQf/ducgsKyMvLA0o+NOle\nD+MXPbHWCwsLfaVH9zM162H8oifI97OwsNBXesLrBQUFjB07FoCcJI+DmsocVQ1gO9AXeNm1/RGg\nDdA1wfP0Bx4D9qEkh1Xk2l8Fe19FWCtsRoSGA4C1wD+BHpTOZUE4R7VokfWdatvW+lIJIYSISabk\nqHZiBRGRpeUnY9V/iZKLFUcArAaOANq5ljFYi6kd8HEUDWuxcvXewOsxrxLuQ6WKPyGESCupLk8f\nCeQDl2Bl4qOxPNEYZ/8ISreA+gP9HNuWWIhvICXVe7uxcJ57+RErnFiKteAAOmI5qWbACcA0Z/u9\nMZUqR+UJ0ukt0uktQdAZBI3JJtU5qslY6G0Y0BjrmNsDWOXsz8acSZiQY9sEC+V9CQwAJsa5Roiy\nRRW1gDucc28D3sYKMrbEPIv6UAkhhC/QWH9lsRzVuHE2H9WFF8L48enWJIQQviZTclTBQi0qIYTw\nBXJUsfB5MUVQ4tbS6S3S6S1B0BkEjclGjioWPi+mEEKIyoJyVGWxHNWQITByJNx3Hwwdmm5NQgjh\na5SjSgfKUQkhhC+Qo4qFclSeIJ3eIp3eEgSdQdCYbOSoYqEWlRBC+ALlqMpiOaqTToKZM2H6dDj5\n5HRrEkIIX6McVTpQi0oIIXyBHFUslKPyBOn0Fun0liDoDILGZCNHFQu1qIQQwhcoR1UWy1E1bAgb\nNsC6dXDQQenWJIQQvkY5qnSgFpUQQvgCOapoFBXBr79CVhbUrp1uNVEJStxaOr1FOr0lCDqDoDHZ\nyFFFw11IkaXoqBBCpBM9hcsSCn3/Pfzud9C4MaxZU/4RQghRycm0HNVA4BtgBzAf6BzHNgcojrJ0\nj2HfGZuefnGUfUOwGYJ/wWYUfhiIPjS6Rk4XQgjfkGpHdT4wCrgTyAU+AqYCvy/nuFOwaerDy7tR\nbPYHxgEzKDsV/cXAXcDtQCtnvQcwOurVAlBIEZS4tXR6i3R6SxB0BkFjskm1o/ob8AzwFNa6+Suw\nFriynOM2Aetdy64oNk855/6Yss3PjsAcYALwHeboxgPHRL2azzv7CiFEZSKVOaoawHagL/Cya/vD\nwBFAXpRjcoCvsVBdLeAr4IGI48HCif2ALsBwoDfQ1rX/dOA54FRgLnAo8DrwDvD3iHOFQlOmQI8e\ncOqpMHVqRd6jEEJUSpKZo6qWjJPG4ECgKrAuYvt6LJwXja1YbulDLPfUE5gE9MdaR2AO6RasdRQZ\n8gvzNnAzMNtZr4aFCSOdlKEclRBC+Aa/l6dvxFpQnwCfYq2lMcD1zv6amOMaCnwb5zznAHcDVwBH\nAr2ArsBtUa2Vo/IM6fQW6fSWIOgMgsZkk8oW1QagCGgUsb0RlqdKlHnAn53XjbHiiGecBcz5ZmF5\nrNOw4oq/Yzmspx2bz7GKvycxZ1XsvkD+mDHkACxdSv1Ro8jNzSUvLw8o+dCkez2MX/TEWi8sLPSV\nHt3P1KyH8YueIN/PwsJCX+kJrxcUFDB27FgAcnJySCap7kc1B1gEXO7atgx4EQvNJcIDwJnAYZij\nbRGxfxBwMnA21srajrXI3gOuc9n1w5zbPpR2VKHQiBFw441www1wzz0JyhJCiMpLpuSoAEZi1Xaf\nYKXpV2D5qTHO/hFAB+AkZ70/sBMoxJzJmVjhRDj0txtYGnGNH4HfIra/BtyA9dv6BHNydwBvEtGa\nAgIR+hNCiMpCqnNUk4HBwDBgIXAc1p9plbM/G2jmsg85tvMwB3MeMIBY/Z9KjoksqvgncC/mnD7H\nQn7TgL9EPUMAiikiQyx+RTq9RTq9JQg6g6Ax2aS6RQXwmLNEY0DE+jhnqQi3UbZIogjr8HtXQmdQ\ni0oIIXyDxvorSyjUty+88AJMnAj9+qVbjxBC+J5k5qjkqMoSCq1dCz/+CIccAvvvn249QgjhezJt\nUFr/k50Nbdv62kkFJW4tnd4ind4SBJ1B0Jhs5KiEEEL4GoX+yhIKhWKNxCSEECIaCv0JIYSotMhR\nBZSgxK2l01uk01uCoDMIGpONHJUQQghfoxxVWZSjEkKICqIclRBCiEqLHFVACUrcWjq9RTq9JQg6\ng6Ax2chRCSGE8DXKUZVFOSohhKggylEJIYSotMhRBZSgxK2l01uk01uCoDMIGpONHJUQQghfoxxV\nWZSjEkKICpJpOaqBwDfADmA+0DmObQ5QHGXpHsO+M7AbWByxvSDGeZbsgX4hhBApJNWO6nxgFHAn\nkAt8BEwFfl/OcacA2a7l3Sg2+2PT1s8AIptE50QcnwNsBSbtwXvwBUGJW0unt0intwRBZxA0JptU\nO6q/Ac8ATwFfAn8F1gJXlnPcJmC9a9kVxeYp59wfU7b5uTni+BOAfYCn9+RN+IHCwsJ0S0gI6fQW\n6fSWIOgMgsZkk0pHVQNoD0yP2D4dOK6cY18B1gEfAL2j7B8INMRaaonESC/FWnLfJ2DrS3766ad0\nS0gI6fQW6fSWIOgMgsZkUy2F1zoQqIo5HDfrsXBcNLYCQ4APsdxTTyxc1x+Y4Ni0BW4BjqFsyC8a\nLYAuzrmEEEL4nFQ6qj1hI/CAa/1T4ADgesxR1cQc11Dg2wTPeSmwBnjbO5mpZ+XKlemWkBDS6S3S\n6S1B0BkEjZlEDSy3FBm6e4ToxRGx6A/84rzOwar3drmWIte2k6JoWA/cEef8y7GWmRYtWrRoSXxZ\nToYwB3g8Ytsy4K4KnOMBSm5INaBNxPKIc842QJ2IY8/DHFlORUQLIYSoPJwH/AZcArQGRgNbKClP\nH4GVl4fpD/RzbFtiIb7fgGviXONWyvajCjMDeGfPpAshhKgsXIl1+P0VmEfpDr/PAF+71i8GPge2\nAT8DnwB/Kuf8w4HPomxvhrWmzt0j1UIIIYQQQggh4lOR4Z285kashfkzVvDxBnB4FLtbsf5fv2BF\nKG0i9tcEHgJ+xFqirwO/S4pi40aseOUhH+psDDyL3c8dWOu8i890VgPuxiIJO5y/d2BdOdKlswv2\n+VuN/W/7R7HxQs/+wHjgJ2cZB9TzSGc14J/AIuf6a7BK4chRcNKtM5LHHZshPtXZAuvXuhnYDiwA\nWqVYZ6XmfGAnlj9rCTyI9eMqb3gnr5iGfTDaAEdgH4a12D81zA1YTu8czIlNwh4WdV02jznbTgSO\nxB4iC0lO5+5O2IO1ELtfftJZ39E2FjgaaAJ0pfSXyg86b8G6YZwOHAqc6awPS6PO07DO872xh9HF\nEfu90jMVyycfg32WlmAPSi901sMGE+gDNAc6AO9jP1bcPwLSrdPNuVgXnNXYKD5u/KCzKeaA7sOG\nwMsBTgUOSbHOSs1colck3p0GLWAVi7uxBxjYiBtrsRZMmFrYA+MyZ70eVmzSz2VzCJabizWQ755S\nD6u+/D/swxh2VH7ReTcwO85+v+h8E8vNunnW2e4HnVsp/cDySk9r7Ff7sS6b451tLTzQGY3wNcOR\nCj/pbII5qJZYVMftqPyicyLWEopF0nRqPipjb4Z3Shb7Yf+fzc56U6ARpTX+iv1KDGs8CqgeYbMa\n+ALv38e/gReB9yg9bJVfdJ6NFd9MwkZDWQgM8qHOqUA37AEF1qLuSkmHdL/oDLO3esIPqGOx0NDH\nLpuPsF/y7oeYl4TDS+HvlF90VgOex0K+X0bZ7wedVYAznGtOw8Lpn2CV3EnXKUdl7MnwTslmNPZw\nDf9DwzriaczGfr1sjLBZhz1cvOJSrIoyHJ4Kufb5RWczLOe4HPs1Nxq4hxJn5Redj2K5ky+w0PMS\nLFw5xmc6w+ytHrfNjxH7QyTvO1cD+BcWYlrjM523OeeLjOiE8YPOg7DQ7k2YozoJc64TgB7J1un3\nIZQqKyOxX6edKe0EYpGIjVe0xDpod8Y+lGAtqkQGA06lzirYL76bnfVFWK5iENYpPB6p1PlXYADQ\nF8ufHIk51ZWUP7p/KnUmQnl60jVRazXgOSxKcUYC9qnUmYflpnP3QEMqdYYbNa9hUzWBdQM6GrgK\nmBLn2L3WqRaVsQF76Eb++myExeNTyQNYYUc37GEV5geXJjeNXPt+wFqGB0TYZLts9pZjsRbo55QM\nW9UFa73sxO6lH3SuAZZGbPsvVrAQ1hDW5SbVOm/G8mmTsXv6HPZDJZwD8ovOMHujJ9KmYcT+LOyX\nu5eaw2G1I7AE/2bXPj/o/D+sOnUtJd+nJljF4nc+0rkBy5mX951Kik45KmMnVmYZmXg+GYufporR\nlDipZRH7vsH+kW6NtbCWTVjjAuyD7rY5BKt08+p9vIp96ds5Sy5Wyv+88/orn+j8kNIVfmDJ2pXO\na7/czywskeymmJJfoX7RGcYrPR9joSR3XuJYrIjIK83VsRzlEVjeb33Efj/ofBSbAcL9fVqD/Vg5\n0Uc6d2LdZ+J9p/ygM+Mpb3inZPMI1oeqK6VnI3aPV3g91u/gHOzL9wKWrHTbPAqsonR56KckN0xQ\nQOl+VH7QeTT25boJOAwrU/6J0pN0+kHnv53z98DKfc/BHqj3pVFnHeyBmYsluf/hvA5/F7zSMwUL\nH3XCHlaLsX43XuisioWpVjvXd3+navlIZzQiq/78orMn9oy8FPtOXYp9x05Lsc5KT7zhnZJNMSUj\nv7uXWyLshmO/uHYQvaNlDaxUfAP2YUt2h18oXZ4exg86e2B9vHZgIYqrotikW2cd4H7sc/cLsALr\ny1IjjTrzKPn8uT+T7pyZF3rqY+XOPzvLOCyP5IXOJlG2hxd32XW6dUYjmqPyi87+WGXiL9h36/w0\n6BRCCCGEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIISoP+dgkdUKIGGhQWiGEEEII\nQRdgDtZ6+gmYi82NFWtsxwux8Sa3YBPPTQYOjjjn6di4a+Hx9s53znGoy+Y4bBbm7dgArY8C+3r6\nzoQQQgSeatg8SPdiU7m3wCZK/CM2ceI2bD6egygZgXwAcCo2onoHYBbmcMIcio1kfT82IWRv4Fts\nMNGwo2qLOcZrgT8AHbGpFF70/B0KIYQINA2wlk6XKPvySSxH1co5R7hVNQKbaNHNjZRuUY0Dnoyw\nyXVsDkzgmkL4AuWohEg+m4CxwDvAW1gLp7x5ztpjUySsxMJ/85ztYSfUyrUtzCcR60dhIcStruUD\nbMr4P1TsLQiRPuSohEgNfwaOAd4HzsJyS5EzSoepgzm1bZijORoLA0LJPFUhyp8UMQt4gpLZY9th\n4cbmwKI9eRNCCCEqD1OACUA/bBI6N0dh4bkmrm29KB0+vJvyQ3/PYUUWQgghRFxygHuwabebAF2x\nCrybnG3FwElY3qg20BCr5LsfaIZV931OaUd1KDYT9X1AS8yRfYMVU4TDim2xar/HsGnBDwPOAMYk\n6X0KIYQIKAcBL2PO6VesOu8eoKqz/1HgR0qXp58HLMcc1hwsTFhE6YIMd3n6e1ilYDHm6MIcBUzF\npvzeBnwG3OrhexNCCCES5hqsDF4IIYTwBYOwvlFNsVzXj8ADaVUkhBBCuBiJhRN3AF8Bt2Odi4UQ\nQgghhBBCCCGEEEIIIYQQQgghhBBCCCGEEEIIIURp/h+07yc5mLFbuQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1078ad10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "report.learning_curve(RocAuc())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('tt', 0.5544309400521632)])"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "report.compute_metric(RocAuc())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<rep.plotting.FunctionsPlot at 0xe591790>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEiCAYAAADTSFSPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XmcjXX/x/HXkCV7JEvITugOkSVpiBaRopKI0a5FqrtF\noVJSaSG6RclIC9q1ScpoUUqlusmS+GmXG9k1OL8/Ptd0nRlnZs45c865zvJ+Ph7nYa7rXOec73yb\nrs/5bp8viIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIiIgklkzgoPPIBn4GZgI1AlzbDpgH/A/YC3wP\njAZKBbi2JTAH+M25di0wA2gR0dKLiIjnZgDvAkcBNYHuwEbgvTzXnQ38DUzHgkRt4EIs8CwGSvhd\n2xPYB7wBnAocA5wA3Ae8GaXfIz+HxfjzRERSTibWyvD3MLDD77gM8CfwSoDXtwIOAP/Oc+1r+Xxe\nhULKcxPWetkL/IQFH4C6WKupdZ7rDwJ98lxzIfABsBu4zvm3Z57XnYYFxiOd46OB2cAW5/Em0NDv\n+trA61hrbBfWGutXyO8iIpISMrGWQ476wArsRpzjXOwG3T6f91gAfBXktQUZB2wFMoB6QFvgSue5\nugQfSNY7547BAsQc4IU8r5uJ2zoqA6wBnsa63hoDTwIbgNLONW9gLbfjnPc93XmIiKS8TGxsZAf2\nzf0gdtOs7HfNrc75ivm8x0TsWzrALYVcm59ywB7ginyer0vwgeSGPNf0cspXzjk+HPgLa7kAXIIF\nEn/Fgc3Aec7xN9h4kEhEFPO6ACIRthg4HjgRmAScAlQL8T18zr9pYZahGTZo/36Yr/e3LM/xfCxI\nnuscn42VM6f77QSsBbTD77ENqAQ0cK6ZCIwElgD3cGhAEwmJAokkmz3Aj8B/geuxG/FEv+dXO/82\nz+f1zXC/0a/xOxdJB51//QNViUAX4raOcmQDc4EBzvEAbLxnr3NcDFiOBVP/R2NgmnPN01iwmeGc\nXwLcGcbvISKSdDLJPUYC1iI5CLRxjnMG0F8N8PrW2GD7TX7XbsIGpgOplM/58lhAuzKf5w93ytTL\n71xbAndtBWotdMACyrHYjLJufs9dho3NhNIddwvwSwjXi4gkrUwODSRgrZK5fsfnkHv6bx3c6b9Z\n5J5mezZ2s34Tu2HXxW7u91Dw9N/7sRlTGViX0onAVX7PLwE+xlo7HbEuuWADCcAPWMvjZ3K3bA4H\nVjm/R2es5dEZeAh35tZEbHC9vvP7L8ImGYgkhM7Y9Myfsf9JBgfxmuOw/8l2O68bFbXSSaKbwaHT\nfwH6Y9/g6/mda48FnS3kXpBYMsDrW2OB6Hfn2h+A57Auo/ykYQP767BAtBELPjmaYoFkFzb43YlD\nA8kB8g8kdzvPPxTguaOw7qs/nPL+CDyFO+ngMazbbg/W4nqewIs2ReLSmcC9QF/sf6BBhVxfAfuf\ndzb2za0vsB24MYplFBGRBLGDwgPJUGzGiX/aijuwlomIiMSBeJ+11QH4COsayLEAS39xjCclEhGR\nXOI9kFTH+nn9/eH3nIiIeCzeA4mv8EtERMRL8Z5N9HcObXlU83sul5rg+zXqRRIRSTrryJ3YMyTx\n3iL5FDiZ3IPt3bHFU/+X9+JfAV+3bviwpoyvXj18c+bgO3gQn8+XUo8777zT8zLEy0N1obpQXRT8\nwE2fE5ZYB5Ky2AKols5nH4O7HwRYxtSFftc/j60fycRSWvTB5uY/ku8nLFgAb78NzZrB+vXQrx90\n7AhLlkT4V4lvGzZs8LoIcUN14VJduFQXkRPrQNIWS9H9FZbS+m7n57ud56tjq21zbMdaIDWx1cmT\nsAVYj+b7CWlpcOaZ8M03MHUqVKsGn30GJ50E558P69ZF+ncSEUlp4WY3jVc+p5nm2rEDHnwQHn4Y\n9uyBEiXg2mth9GiolF+qpMSXlZVFenq618WIC6oLl+rCpbpwpaWlQRHiQfIHkhw//wyjRsHMmeDz\nwZFHwtixcOmlULx4bEspIhJHihpI4n2wPXJq1YIZM+DLL6FzZ9i8Ga68Ek44ARYv9rp0EZeVleV1\nEeKG6sKlunAVVheVK1cmLS0tqR6VK1cu8HcOV+oEkhytWkFWFsydC3Xq2FhKerqNq3z+udelE5E4\nsXXrVs9nU0X6sXXr1qjUVep0bQWyZw889JCNoezcaed69rQur3/9KzolFJGEkJaWRkj3kwSQ3++k\nMZLcQgskOTZvtoAyaRLs3m0zvy6+GO65x1otIpJyFEiCl3pdW4EceSTcf7+tOxk2DA47DJ55Bho3\nhttvt5lfCUZ94S7VhUt14VJdRI4Cib+jjoKJE2HVKujfH/btg3HjLKBkZsLBg4W+hYhIqlHXVkE+\n+wyuv94dhG/TBiZMsMWNIpLU4r1rq27dujz99NN07dqVDRs2UL9+ffbv30+xYvm3D9S15YX27eHT\nT2HWLKhZE5Ytg06drLWycaPXpRORFBYoKHgV+BRIClOsGAwcCGvW2ILG0qVh9mxo2hTuvht27fK6\nhAGp/9elunCpLlyJXBcXX3wxGzdupFevXpQvX57mzZsDUKlSJcqXL8/SpUtjWh4FkmCVLQtjxtj4\nSb9+NnX4rrugQQOb7bVvX6FvISJJJC0tco8QzZo1izp16vDmm2+yY8cOVq5cCcBff/3Fjh07aNeu\nXaR/2wIF+xsUB9oAjYGK2D7qa4Evgf3RKVpYIjtGUpDFi+Hmm+GLL+y4Th3L3zV4sM36EpGEVugY\nSRgBIF9h3Lfq1avH9OnTE2KMpC2Wwn0btjfIZCyN++PO8Tbn+RPDLUDCOuUUWLoUXn0VWrSwMZPL\nLrP09bNna4aXSLLz+SL3SHAFBZJXgTeBndg+IFWw1kht598qQF9gF/AG8EpUSxqP0tLgnHNg+XJ4\n7jlo2BDWrrXB+I4dLa+XRxK5/zfSVBcu1YUr0esiza9FlBbJ1lEYCgoki7CNp64F3gPyJmnZCrwL\nXAPUBbIiX7wEUbw4XHQRrFwJ06ZBjRrWWmnbFq64Av74w+sSikiSqVatGuuc/ZWqVq1KsWLF/jmO\nNa0jiYYdO2xgfsIE2L8fypWDESPghhvg8MO9Lp2IBCHe15HMmzeP6667ju3btzNy5Ei2b9/OlClT\nyM7O5t133+XEEw8dcfA619brwFPAW0A8d/7HRyDJsWoV3HILvPGGHdeubSvl+/e3acUiErfiPZCE\nw+sFiTuB2cDP2L7qjcL9wJTStCnMmwfvvw8tW8JPP9malA4dor6HfKL3/0aS6sKlunCpLiIn2EAy\nANs3/R6gG7Aa+BAYDKivpjBdu9qq+KefhurVLeXKSSfBhRfChg1el05EpEjCbcq0AC4FhgJ7gTnA\nRGBlhMoVrvjq2gpk50544AFLW793L5QqZWMnI0ZAhQpel05EHOraCl44HfU1gd5ATyAbm/ZbB/gW\nuDncgqSMcuVsn5PVq2HAAFsRf//90KgRPPkkHDjgdQlFREISbCApCZwPvANsxALJg0AN4BLgTGyt\nyR1RKGNyqlMHnn3WMgx36ACbNtlU4VatYOHCIr+9+n9dqguX6sKluoicYAPJr8BU4AegNbaS/Uls\nED7HR9hKdwlFu3bwySe2Gv6YY+C776B7d+jVy1otIiJxLtg+sYuBF7HxkHgW/2MkBdm719ae3Hef\nrUU57DAYOhTuvBOqVPG6dCIppXLlymzdmncddmI74ogj2LJlyyHnY7WOZAYwDMi752xZYBLWvRUP\nEjuQ5PjjD0tZP3265ew64ghLCHn11VCypNelE5EkE6vB9vym+ZZxnpNIqlbNUq18/TWceips3Woz\nu1q0sHUpQQRL9f+6VBcu1YVLdRE5hQWSylhyxpyf/R9VsZlbSiQVLf/6F7z3nq2Mb9LEEkL27g3d\nulmiSBGROFBYU6awdCg+4E7g3sgUp8iSo2srkOxsmDLFNtPautUyDw8aZDm96tTxunQiksCiPUaS\n7vz7AZYy3n/k6W/g/4Bfwv3wKEjeQJJjyxZbhzJ5siWELF3aNti67TYoU8br0olIAor2GEmW86iP\nJW7M8nssIb6CSGqoXBkefdTd8nfvXgssTZrYnijOhlrq/3WpLlyqC5fqInIKCiStsS12wcZEWjrn\nAj0k1ho0sLUnH39sixh//tkSQnbsaIscRURipKCmzEGgOrCJgsdKfLgBx2vJ37UVyIEDMHMm3HEH\n/P67nRs8GMaPh6pVvS2biMS9aI6R1MXSoRx0fi7IhnALEGGpGUhy7Nhh+5088ojl8KpQAUaOhGHD\nLDmkiEgA0Rwj2YDbEtlQyEPiQfnytir+22/JOvFE2L7dNtZq1gxefjmo9SfJSH3hLtWFS3UROcEu\nSLwOGBjg/EDg6hA/82pgPbAHWAZ0KuT6HsBnwHbgT+A1tLFWwRo3tlT18+dbEPnxRzjvPEhPhy+/\n9Lp0IpJkgm3KrMNWsH+c5/zJWPqUhkG+Tz9gFraPycfANcAQoBnwU4DrG2J7nDwCTAPKAw8ADQgc\nTFK7ayuQ/fstPf3o0bB5s50bNMhaLkcf7W3ZRCQuxCrX1l6gKYd2Y9UDvgdKB/k+S4HlwJV+59YA\nLwG3B7j+PGyL3xLYoD5AF+B94Eggb/YxBZL8bNsGY8fCxIm2uLFMGev2uvlmrT8RSXGxyrX1O9Aq\nwPlWwOYg36MkNlV4QZ7zC4CO+bzmEyxV/eXYzLDyQAbwOYcGEfFzSP9vpUo2i+v776FPH9i921bJ\nN24Ms2b9s/4kGakv3KW6cKkuIifYQPI88BhwGtY6KAGcjm2v+1yQ73EkFgzy5ubahE0zDuQ3bIzk\nXqxVtA1oDvQK8jMlrwYNbOB98WJo3Rp++cW6utq1szUpIiIhOizI6+7CurHm487kKgbMBUZFvlj/\nqI8Nrs/AglkFYIzzuV1xu7v+kZGRQd26dQGoVKkSLVu2JD09HXC/gaTCcXp6esHXd+5M1vjxsGAB\n6bNmwbJlZJ18MpxyCukzZkC9enH1++g4csc54qU8Xh3nnIuX8sTyOCsri8zMTIB/7pdFEWqfWCPc\nLq7l2PhGsEoCu4ALgZf9zj+ODbZ3CfCaB4BuwAl+547GBuY7YWla/GmMJBy7dsGDD1rX1549tufJ\nDTfA7bfbWhQRSWqxGiPJsRZrDcwltCACluTxS6x7zF93Dg0IOdI4dFW9f4tI8pH322eBypaFu++2\nrX0HDoS//7bpww0bwtSpNvMrgYVUF0lOdeFSXUROQTfjx7AdEMF2QXwswCPnfLAewQbLLwWOxcZY\nqgNPOM+PAxb6XT8PG6AfhbWGWmPdXBuxoCSRVLu2DbwvXQonnQR//glXXWW5vN57z+vSiUicKqgp\nkwWcgw1wZxFgPMJ5vY/A3VL5GQrcAtQAvgNuwF2fMgM4BRsbyXEecBvQBNgNfOocrwrw3uraihSf\nD158EW69FTZssHNnnQUPPQRNm3paNBGJrGiuIzkGG4tIpHmhCiSRtncvTJhgCxh37IDixWHoUJs6\nXKVKoS8XkfgXzTGSH7Epu2AbW1UK90Mk9iLW/1u6tG2atXYtXHGFtVQmT7bxk0cftfGUOKe+cJfq\nwqW6iJyCAskObF92sJ0SS0a9NBK/qlWzgffly23P+G3b4MYboUULeP31lE0IKSIFN2VewnJpfQ90\nxmZWZQe4zoet6YgH6tqKBZ8P3noL/v1vm+kF0KWLpa9v2dLbsolIyKI5RlIGuARLnDgMGwjfE+A6\nH5YdOB4okMRSdjY88YSNl2zZAmlpcMkltvVvjRpel05EghSrpI1ZwLnA1nA/KEYUSBz+K3ajbssW\nCx6TJ9uak7JlYcQI6/o6/PDYlKEAMa2LOKe6cKkuXLFakJhO/AcR8UrlyjbwvmIFnH22rZQfOdKm\nCb/wgsZPRJJcQRHoMWAEltZkEgWvIxkW+aKFRS2SePD++9Ya+fZbO27f3gJN+/belktEAopm11YW\n0VmQGE0KJPHiwAGYMcNaJn84CZ/794f774c6dbwtm4jkEs2urXQsiOT83CXAI+e8xBnP58gXLw6X\nXWbrT0aMgFKlrJurSRMLLjt3xqwontdFHFFduFQXkRPsGElJINCo6eFofYkUpHx5WxW/ejVceKGt\nlB87Fho1gunTreUiIgkt2KbMPGx1+4Q854djrZJzIlimolDXVrz79FNLUb90qR23bGnrT7qoYSvi\nlVjN2uoIBEr/+h5wUrgfLimoQwdYsgSee86yDS9fDl27wjnnWDeYiCScYANJGSBQH4QP20dd4kxc\n9/8WKwYXXQSrVtn6k7JlLc1K8+Y222trZGeax3VdxJjqwqW6iJxgA8l3wEUBzvcH/hu54khKKVPG\nBt7XrrUV8fv32zThhg1h0iRbOS8icS/YPrEewOvAi8D7zrluwPnYivc3Il+0sGiMJJF9/bW1SHK+\nKTZtavuf9Ohh6VdEJCpilSIF4Axsp8KcrHxfA2OBd8L98ChQIEl0Pp91c918M/zwg53r3t0G5Fu0\n8LZsIkkqlnu2z8cG1ss6j07EVxARPwnb/5uWZgPvK1bAww9DxYq2ze/xx9u2v5s2hfyWCVsXUaC6\ncKkuIieUQHI41pV1K+4mVw0BbZMnkVeypHVz/fADXHONBZipU239yYMPwr59XpdQRBzBNmUaAguB\nclgQaYztoPiQc3xZVEoXOnVtJauVK23/k3ecRnDjxpZu5ZxzNH4iUkSx6tqagK0ZqUbuPUnmET+b\nWkkya9YM3n4b5s+3ILJmDfTpY+Mnq1Z5XTqRlBbKgsTxHLqW5CegZkRLJBGRtP2/p58O//2vTQ+u\nVMkyDbdqZfm8duwI+JKkrYswqC5cqovICWWMJFBOrdrAXxEqi0hwSpSAa6+FdesgI8Pyd91/v9af\niHgk2D6x2cBubOvdHcDxwP+wtSU/OufjgcZIUtHSpXD99W7+rmOPtQH5s87S+IlIEGK1juRoYJHz\ncz1gOTYA/wfQGQh9TmZ0KJCkqpz1JzfdBD/+aOd69rQFjU2aeFs2kTgXq8H2X7CFiPcD04BlwM1A\nK+IniIiflOv/zVl/snKlLV6sUAHefBNatCDr/PNh27bC3yMFpNzfRQFUF5ETyhjJbuBp4BpgKPAU\nuWdwiXivVClLU796tW2sdeAAvPSSrT958kk4eNDrEooknVCaMidg+480w7L+fo9NC/4yCuUKl7q2\nJLevv7bxk48+suO2bW1Avl07b8slEkdi1bU1APgcqA68jaVGqe6cuzjcDxeJulatYPFimD0bjj4a\nvvgC2reHIUPgzz+9Lp1IUgg2kIzFEjZ2d/7N+XkkcE90iiZFof5fV9bixdCvny1cvO02S7+SmWnd\nXdOmpVR3l/4uXKqLyAk2kFQF5gY4/xJwVOSKIxJF5crBuHG2oPH00+Gvv+DKK23Xxs8+87p0Igkr\n2D6xt4DXgCfznL8M6AucGclCFYHGSCQ4Ph/MnWsD87/9ZucGDLCFjbVqeVs2kRiL1TqSa4C7gZeB\nT51zHbBNre4Cfve79pVwCxMBCiQSmp07LXg89JBlFC5bFu6+G667zrrARFJArAJJKJ3IoUwpjjQF\nEkdWVhbp6eleFyMuBFUXGzZYduGXX7bjBg1g/Pikyy6svwuX6sIVq1lbxUJ4FOZqYD22BmUZtkFW\nYYYDq4C9wK/AuCDLLRKcunVtvclbb9kWv+vWWXbhM8+0ICMi+SrKV60SQKjZ8foBs7AFjR9jXWZD\nsLUpP+XzmkeAs4B/A98BFYEa2I6NealFIkWXnW2zuUaOtBXxZcrAHXfYepSyZb0unUjExapr63os\nTcpLzvHTwGBgHdALWB3k+yzF8nRd6XdujfO+twe4vgkWPI4L8jMUSCRyNm2ysZK5zoTFunXh8ceh\nRw9PiyUSabHq2hoG5Kze6oxtuXsR8DXwcJDvURJoDSzIc34Btt9JIL2x7MI9nH/XA5nYdGQpgObI\nu8Kui6OOgjlzYOFC2zN+wwbLKHzJJbBlSySLGDP6u3CpLiIn2EBSE7uRg7VAXgLmYDO2OgT5HkcC\nxbGMwf42YavkA6kPHANcAAzCVtE3Bd6gaN1yIsE79VRYtswG30uVghkzbDB+4kTYv9/r0ol4Ltib\n8R/YOMUyrGvqIeBZoJFzHEzHcU3gZ6xF87Hf+dFY66ZpgNdMw9aqNAZ+cM41wrq52gFf5LleXVsS\nXatXwzXX2M6MYFsA/+c/cMop3pZLpAiK2rV1WJDXLcAWI36F7UPyjnO+GdbdFIzN2Fa91fKcrwb8\nls9rfgP24wYRnJ8PAHU4NJCQkZFB3bp1AahUqRItW7b8Z4pfTlNWxzou0vHChTBvHllXXQUrV9rz\nAweS1aMH1Kjhffl0rONCjrOyssjMzAT4534ZCxWBSdiOiGf4nR9D4EHy/HwGTM1zbg2WyyuQ7tga\nlvp+5xo459oEuN4nZtGiRV4XIW5ErS727vX57rrL5ytVyucDn690aZ9vzBifb8+e6HxeBOjvwqW6\ncGEZ3cMW7BjJX8B12OC3/7Tb0cB9IXzeI0AGcClwLDARGx95wnl+HLDQ7/qFWCvoaWxjrVbOz59h\n3Wwi3ilVCu68E1asgP79be/40aPhuOPg3Xe9Lp1IzBTUJ1YB2B7CewV7/VDgFmwtyHfADbhjJjOA\nU8jdAqkOPIa1hPZg3Ww34s4i8+cEVxEPZGXZ+MnKlXbcpw88+ijUqeNpsUQKE811JJuwwe4Z2HqR\n/DTEFhVejveZgBVIxFvZ2Tab6667YNcuW8w4ahTceKNyd0nciuY6kvZAXWAlsAJbvzEGa02MAWb6\nPXeMc73EiZyBNYlxXZQoYTm7Vq2CCy6A3bthxAho2dLdpdFD+rtwqS4ip6BA8iMwEKgNPA6UxhYG\nXoaljS8JTMZmTw3EXWciIrVq2WLGBQugcWP4/nvo3BkuusiCjEgSSbZFferakvizdy/cd5+lq8/O\nhsMOs66uUaNssy0Rj8Uq11aiUCCR+PV//wf33gvTp9vGWkcfbeMpffokVap6STyxyrUlCUb9v664\nqYtjjoEnn7Rtfdu0gV9+gfPOg759LUFkDMRNXcQB1UXkKJCIxNqJJ1owefxx69p69VUbR5k0Sbm7\nJCElW3taXVuSWDZuhCuucBcwNm1q2/726KHuLomZaHZtPQ2Ud37ujG1kJSKRVKcOvPMOvPKKZRRe\ntQp69oR+/WLW3SVSVAUFkouBnCklWcARUS+NRIz6f11xXxdpaXDuuZZq5eGHrbvrxRehSRPr7jpw\nIGIfFfd1EUOqi8gpKJBswPJrpTvHHbGWSaCHiBRVqVI2Lfibb+D0022b32HDoH17+Oorr0snkq+C\n+sR6A9OByoW8hw/bsCoeaIxEkoPPB/Pm2Va/P/0ExYrZnvFjxmjtiURcLNaRHAH8D2hO4ESJYHuN\nxAMFEkkuO3daRuGJE+HgQahdGyZPhrPP9rpkkkRiMdi+FeiCbSi1OZ+HxBn1/7oSui7KlYNHHoEv\nvoATTrDWSe/eNhj/008hv11C10WEqS4iJ9jB9g/QYLuId1q3hqVLYcIEyyg8dy40agT33AP79nld\nOklxBTVl1gIvYvt/fAD0Abbkc+2HES5XuNS1Jclvwwa49VYLJmBrT554QvvGS9iiOUaiwXaReLZo\nEVx1FaxZY8dDhsD48VClirflkoQTzTGS14EjgZy/yubYxlV5H9XC/XCJHvX/upK2Lrp0gW+/tU20\nSpaEGTOsdTJzps36CiBp6yIMqovICSbX1lagKxpsF4k/OfvGf/utBZbNmyEjA7p2hdWrvS6dpIhQ\nmjLVsQH4+sAoLIB0An4B1ke+aGFR15akLp8PZs2Cm26ygFKypO3OeNttULq016WTOBar/UhOwAbc\nf8S6uJo6P98NNAIuCrcAEaZAIvK//8Ett8DTT9tx48Y2GN+li7flkrgVq/1IHgYmAq0A/7mG87FW\nicQZ9f+6Uq4uqlSxzbMWL7YxkzVrrKtr8GCyXnvN69LFjZT7u4iiYANJayAzwPnf0WC7SHzq3BmW\nL7e1JqVKwTPPwKBBNiivlrtEULBNmT+As4BlwA7geKxr6wzgSaB2VEoXOnVtiQSydi1cfTUsXGjH\nnTtbd9exx3pbLokLseraeh24E/AfsasHPAi8HO6Hi0iMNGoECxbAs89C1arw4Ydw/PGWx2vvXq9L\nJwku2EByM5Yi5U+gDPAxNh14GzAyOkWTolD/r0t14UhLI+voo23zrMsvh+xs6/Y67ji3pZJC9HcR\nOcEGkr+Ak7HV7rdhA++nY3uR7IxO0UQkKipXhmnT4KOPoFkz+OEH6N4dBg7UrowSlmTbFFpjJCKh\n+Ptv25VxzBjr4jriCEtT37+/9oxPIbFaR5IoFEhEwrFunQ3GL1hgx6edZoPx9ep5Wy6JiVgNtkuC\nUf+vS3XhyrcuGjSA+fPhySetVbJgAbRoYa2V/ftjWsZY0d9F5CiQiIhJS4PLLrPB+H79YPdu+Pe/\noUMHWLHC69JJHFPXlogE9vbbMHQobNwIxYvDDTdYpuGyZb0umUSYF2MklTi0JZPfhlexpkAiEkl/\n/WVJH6dOtdXwderApEnaMz7JxGqMpC6WV2svFjT8U8j/Ge6HS/So/9elunCFXBcVK8KUKbbNb6tW\n1jrp3Rv69IFffolKGWNFfxeRE2wgeRrbxOoS4FRsf5Kcx6nRKZqIxI22beHzz23P+HLl4NVXoXlz\na50cOOB16cRjwTZldgIdgO8i8JlXYyvlqwMrgOHYSvnCNAK+cn4un8816toSibaffrKpwm++aced\nO8Pzz8PRR3tbLglbrLq2NgClwv0QP/2ACcC9QEtgCfAOhSd9LAnMBhZje8SLiFdq14Z586xVUr26\n5e1q2RLefdfrkolHgg0kw4D7sFZBUdwIzACmA6ud9/0NGFrI6x4AlgMvknwzzaJC/b8u1YUrYnWR\nlgbnnGNb/J52mu3IeMYZts3v5sTYfVt/F5ETSvbfdOzmvwdLJZ/z2B7ke5TE9jVZkOf8AqBjAa87\ny3lch4KISHypWhXeeQfGjrU9T2bOtM20nnvO65JJDAV7Y84o5PnMIN6jJvAzlujRf0xkNLZVb9N8\nXvMFcI7ihKJeAAAaOUlEQVTzbwYwCY2RiMSfNWvgqqtg0SI77tMHJk6EWrW8LZcUqqhjJIcFeV1m\nuB9QRLOAKVgQEZF41rgxvP++bfN7443wyivw3nvw4INwxRVQTIk0klWwgQRsU6sBwLHYgPdK4Hly\n7+FekM3AAQ7dmrcaNk4SSBesBXOnc5yGdcdlY+MqT+V9QUZGBnXr1gWgUqVKtGzZkvT0dMDtE02F\nY//+33goj5fHOefipTxeHi9fvpzhw4dH9/Muuwy6dydrwAD45BPShw6F2bPJuuIKqFkzbupjwoQJ\nKX1/yMzMBPjnfhkLzYCN2EZWH2FdU9uA/8MCS7A+A6bmObcGGFvA5/o/bgd2OT9XCnC9T8yiRYu8\nLkLcUF24YloXBw/6fHPn+nxHHeXzgc9XtqzPN3myz3fgQOzKUAD9Xbgo4mzYYPvE3gN2AxfjDq5X\nAJ7FWiqnBfk+F2DdVVdjU3+vAoYAzYGfgHFAW6BbPq/PQGMkIonlf/+Da6+F2bPtuEsX6/5Sivq4\nEat1JCcBd5B7htZ251ynED5vLrYAcSTwNTZbqwcWRMAWKdYv5D0UKUQSSZUq8MIL8NJLNstr0SLb\n3nfKFDh40OvSSQQEG0j2ErgrqaLzXCimAPWwlkxbcs/gGkLBgSQTawlJIfzHB1Kd6sLlaV307Wvp\n6C+4AHbtstXx3bvDhg2eFEd/F5ETbCB5A5iGtT6KO4+TnXPzolM0EUk6VavCnDnw4otw5JHwwQfW\nOpk6Va2TBBZsn9gRWGugF5DzX7sYtlBxCDbwHg80RiKSKP78E665xoIKWM6uqVNtQaPEVKz3I2mE\nO0vre2BtuB8cJQokIolm7lwbjP/zTzj8cLjzTttEq2RJr0uWMmK9Z/tarCtrHvEXRMSP+n9dqgtX\nXNbFBRfY9r6DBsGePbaRVvv28F0kko3nLy7rIkEVtCDxMWAEtm5jEoFnS6U554dFvmgikjIqV7Y8\nXRddZGlWvv4aWreG22+HUaPgsFDWTkusFdSUycJyXG1zfi4okHSJdMHCpK4tkUS3fbsFkMcft+MO\nHWy/kxiuwE41XuzZHs8USESSxeLFMGCAbelboQL85z92LBEXqzGS0UCZAOcPd56TOKP+X5fqwpVQ\ndXHKKfDNN3DuudZKGTgQevWK2LqThKqLOBdsILkLKBfgfFnnORGRyKtSBV5+GZ56ylolb75pYyfz\ntHwtngTblDmIpS/ZlOd8N+AFoGokC1UE6toSSVa//QaXXw5vvWXHV10Fjz4KpUt7W64kEO2urZxd\nEAF+JPfOiLux3Q1fDPfDRUSCVqOGtUQefdR2Y3ziCRuI/0LbFXmtsEBynfMAS+F+nd/jMixlytVR\nK52ETf2/LtWFK+HrolgxGD4cliyxWVzLl0OnTtb1FWJvRMLXRRwpbHJ2pvPveizte3ZUSyMiEozW\nrW3B4i23WBbhyy+33RmnTIFKgfLLSjQF2yd2AbYT4ut5zvcGSgAvRbJQRaAxEpFU88wzlkl41y5r\npcyZAyee6HWpEkqspv/ehY2J5LUbzdoSES8NGmQr4U84waYGd+oE996rbMIxFGwgqYdtiZvXD0CD\nyBVHIkX9vy7VhStp66JRI/jkExg2DLKzLa1Kr16wKe9EU1fS1oUHgg0kW4EmAc43wp3VJSLinVKl\nYOJEePttOOII+/f44y3ASFQF2yc2BdvIqi+w2jnXFHgZ+AS4IvJFC4vGSEQEfv7Z0ql8+KElfHz4\nYbjuOkhLtqxQkRGrXFsVgbeB9sBvzrkawFLgTOCvcAsQYQokImKysy3540MP2XGfPjZN+IgjvC1X\nHIrVYPtf2JqRM7D08o8BpwMnET9BRPyo/9elunClVF2UKAHjx9vGWRUqwCuvQMuW4NRBStVFlIWy\nsZUPeA940HksJHBqeRGR+HH++Tarq21b2LgRunSxRY379nldsqQRbFMmDVvBfjVQH2iOpUy5zfl3\nblRKFzp1bYlIYH//DePG2dTg/fuheXN47jkbkE9xserauh4YCTyZ5/yvwLXhfriISMyULGn7wX/6\nKTRuDCtW2MLF8ePhwAGvS5fQgg0kQ4HLgQnAfr/zXwEtIl0oKTr1/7pUFy7VBdCmDXz1FVm9elkr\n5ZZboGdP2LzZ65IlrGADSR3guwDns7HNrUREEkfZsnDjjba/SeXKMH8+HHfcPwPxEppg+8RWYl1b\nr2ALEI/HxkaGAxcDJ0SldKHTGImIhGbjRtt98aOPoHhxmy58/fUpteYkVmMk44HJwADnNR2xHFv3\nOc+JiCSmOnVg0SK49VYbK7nhBrjwQtveV4ISbCCZAdwJjMO6sp7B9iO5DpgdnaJJUagv3KW6cKku\nXLnqonhxuP9+eOklKFfO1p506GB7xkuhQllH8iQ2VlINW9VeC5gejUKJiHiib1/48kto2hRWroR2\n7WDSpJA3zUo1ydYJqDESESm63btt0eKTzoqHq66Cxx6z1fJJKJq5tr4DOmOZfwPN2PK307nmHuCn\ncAsTAQokIhI5c+bA4MG2Cr5jRzuuVcvrUkVcNAfbXwb+9vu5oEcW0BJ4LtyCSGSpL9ylunCpLlxB\n1UW/fvDBBxY8liyBVq1gwYKoly3RFLRn+135/JyfBsCKohRGRCTudOxouboGDLAgcsYZMHq0bZ5V\nvLjXpYsLoTRl0rD1Ig2At7DurHLYXu7ZzjWVgG2RLGCI1LUlItFx4ACMHQt33WWD7926Wa6uo47y\numRFFqt1JNWAT4HPgeeBnJp7GHjI77pgg8jVwHpgD7AMS1Gfn3TgdSyv1y7gG2BIkJ8jIhIZxYtb\nS2TBAqhaFRYutISP6uoKOpA8CmwCqgC7/c6/iO1LEop+WM6ue7FxlSXAO0DtfK7vgAWPvljW4SnA\nNKB/iJ+bUtQX7lJduFQXrrDrols36+o6+WT4/Xc480x45JGUniIcbCA5Fbgdm8Hl70dsbUkobsQW\nOE7Htu0dhu26ODSf68cBo7EW0QbgCSxVS98QP1dEJDKOPtpWw48eDQcPwk03Qe/e8OefXpfME8H2\niW0H2mI3fv9cWycC84HKQb5PSax76kJstleOyVgW4fQg32c+sJFD94rXGImIxNYrr8Cll8K2bVCv\nnrsTYwKJ1RjJR0BGnnOHAbcC74fweUcCxYE/8pzfBFQP8j16Al2x7i0REW/16QPffWfp6devh5NO\ngmef9bpUMVXQ9F9/NwMfYq2SUtgAewugIrZve6ychK1VuQ4bpD9ERkYGdevWBaBSpUq0bNmS9PR0\nwO0TTYVj//7feCiPl8c55+KlPF4eL1++nOHDh8dNebw8njBhQuTuD7VqkXXvvTBhAunz58PFF5P1\n7rswZAjpXbvGxe+b9/6QmZkJ8M/9MlZqAGOwqb9vY4PlNUJ8j5LYVOG84xuPA4sKeW0n4C9sTCU/\nPjGLFi3yughxQ3XhUl24olYXkyf7fMWK+Xzg8/Xp4/Nt2xadz4kgoEhjAsH0iZXEurYGYWMkRfUZ\nNgvrSr9za7AZYHfk85rOwJvYoPuEAt7bqRMREQ/Nnw8XXAA7dti2vvPn2/hJnIrFGMnfQD2KGLH8\nPIKNt1wKHAtMxMZHnnCeHwcs9Ls+HZsePAV4wbm2OlA1QuUREYmsM86Ar76yXRfXrLHxk48/9rpU\nURPsYPsz2J7tkTAX21lxJPA1tklWD9xkj9WB+n7XDwZKY+M0v2ELE38FlkaoPEnJf3wg1akuXKoL\nV9TromFD23XxzDNhyxbo2hWeeiop15sEO9heBhgIdAe+xKbwgjWFfBQ8bhHIFOcRSN5V60MCnBMR\niX8VK8K8ebbO5LHH4PLLYdky2+MkiVLSB9snluX3s384zQkkXSJVoCLSGImIxKeZM+HKKy0l/ckn\n2y6M1YNd9RBd0dyPJBEpkIhI/Fq61Nad/PorNGgAb79tg/Eei9WCREkw6gt3qS5cqguXJ3XRrp1t\n5XvCCbBunR0nwX8TBRIRkViqXt2Cx9lnW1qVM86A117zulRFoq4tEREvHDgAw4bBf/4DxYrZjK4h\n3swrUteWiEgiKl4cJk92Mwhfcgk8/LDXpQqLAkmSUl+4S3XhUl244qIu0tLg7rth4kQ7/ve/YcSI\nhFtrEuw6ks75nPcBe4F1wJaIlEhEJNUMGwZVqsDgwXD//bB5MzzxRMLsCR9sn9hBLGjkd70PmIct\nWtyVzzWxoDESEUlcb70F550He/dCv34wa1ZMFi7GaoykB/A9MABo5DwGACuA84A+2La5D4RbEBGR\nlHfWWbYHfPnyMGcOnHsu7NnjdakKFWwguRfLj/UC1o21zvn5Bixj7+vYHiE9o1BGCUNc9P/GCdWF\nS3Xhitu6OPlk+OAD6+p66y3L1bV9u9elKlCwgaQ58EuA8786zwH8l+B3ORQRkfy0aQMffgg1a8Li\nxXDqqTZuEqeC7RP7CliJpX7f55wrDTyJBZLWwMlYlmAvk+5rjEREksf69dCtG/z4IzRrBu+9Z8El\nwmKVa6sdtrFUMeA753UtgANALyyl+2DgKGB8uIWJAAUSEUkuv/4K3bvDypW2OdbChVC/fuGvC0Gs\nBtuXYi2N27E9RL4CRmD7huTsCzITb4OI+Inb/l8PqC5cqgtXwtRFzZrWzdWmjbVQOnWCFSu8LlUu\nwa4jAdgJTI1WQUREJB9VqsD771t+rsWL4ZRTbPveNm28LhkQfFOmFO7YSC1st8QywBvAh1EoV7jU\ntSUiyWvPHjj/fJvNVb48vPGGBZUiivYYSRPgVaAp8C22dmQBUMF5/nDgfOeaeKBAIiLJLTsbBg2C\n2bOhdGl4+WXo0aNIbxntMZKHsCm+Z2OD7G8D7wIVncdU4NZwP1yiJ2H6f2NAdeFSXbgSti5KlIBn\nn4UrrrAV8L172+JFDxU2RtIeOA0bYF8M/AX8B0uZAjAZ+CxqpRMRkUMVL265uCpWhPHjoX9/S0t/\n0UWeFKewpsxBbJHhJud4B3A88KNzXB1rscRLFmF1bYlI6vD5YOxYGDXKWirz50PXriG/jdf7keiu\nLSLilbQ0GDnSsgdnZ9usriVLYl6MYALJLCyz7xvYavZpzs/znOckDiVs/28UqC5cqgtXUtXFo4/C\nwIGwaxf07AnffRfTjy9sjOQZcqePfy7ANTMjWiIREQlNsWIwYwbs2AGvv25pVT76CBo3jsnHa892\nEZFksXcv9OplaVSOOQY+/RRq1Cj0ZbHKtZUoFEhEJLXt2mXZgpcuheOPt/QqFSoU+BKvB9slTiVV\n/28RqS5cqgtX0tZF2bK24r1RI/jmG+jTB/7+O6ofqUAiIpJsqlaFd9+FatUsR9eAAVENJuraEhFJ\nVl99BV262A6Ll10G06bZlOE81LUlIiKBtW5te8CXLg1PPQV33RWVj1EgSVJJ2/8bBtWFS3XhSpm6\naNfOEjwWKwZjxsDjj0f8IxRIRESSXe/e1iIBWwW/cGFE315jJCIiqeKOO+C++6B6dfj2WxuUJzHH\nSK4G1gN7gGVAp0KuPw7LPLwb+BkYFdXSiYgkqzFjoHNn+P13GD48Ym8b60DSD5gA3Au0BJYA7wC1\n87m+AvAe8BvQBrgeuBm4MeolTXAp0/8bBNWFS3XhSsm6KF4cMjNt8P355yPWxRXrQHIjMAOYDqwG\nhmFBYmg+1w/AEkUOBlYCLwMPoEBSqOXLl3tdhLihunCpLlwpWxf16sHo0fbz1VdbWpUiimUgKQm0\nxrbq9bcA6JjPazoAH+HuF59zfU3gmEgXMJls27bN6yLEDdWFS3XhSum6uOkmaNYM1q6FceOK/Hax\nDCRHAsWBP/Kc34RtkBVI9QDX/+H3nIiIhKpkSZg61X5+4IEiv128T//VFKwwbdiwwesixA3VhUt1\n4Ur5uujUyTbC2rev8GsLEcvpvyWBXcCF2FhHjseBZkCXAK+ZCVQBevqdawssBeoB/5fn+h+ABhEq\nr4hIqlgHNPS6EMH6DJia59waYGw+118F/AWU8jt3O/BT5IsmIiKJ4AJs4PxS4FhgIrAdd/rvOMB/\nPloFbFbXC0BzoA8WWG6IUXlFRCQODcUWJO4FviD3gsQZwI95rm+BLUjcA/yCFiSKiIiIiIhXlF7F\nFUpdpAOvA79iEx6+AYZEuXyxEurfRI5GwA7nkSzCqYvhwCqsh+BXrHs5GYRaFz2wMdztwJ/Aa9jf\nSKLrDMzD7n8HscXdhUnm+yb9gL+x8ZUmwGPYTaCg9Cq/A7OxWWF9sT+SZFgVH2pdjADGYAs862KT\nGLKB/tEuaJSFWg85SgJfAm9ifxPJIJy6eATLMNEL+7s4HjgjqqWMjVDroqFz/f1Afawe5gNro17S\n6DsTS0nVF/sSOaiQ65P5vgnYlN9AM77uy+f6ocA2cs/4ugOLsIku1LoIZA7wUsRK5I1w6+FRLE3P\nYJKnRRJqXTTBbp5Nolkoj4RaF+cB+8m9HKIL9g2+csRL550dFB5IwrpvxvuCxBxKr+IKpy4CqQhs\niVShPBBuPZzlPK4jebZRCKcuemMTW3o4/64HMoGq0SlizIRTF58AO4HLsewb5YEM4HMS+/+RcIR1\n30yUQKL0Kq5w6iKvnkBXYFoEyxVr4dRDTex3HoD1/yaLcOqiPnZjuAD7lnox0BR4g8QOsOHUxW9Y\nQL0XGyvahi036BWlMsazsO6biRJIwqH0KoGdBDyHfSNf5nFZYm0WMAWbdp7qimHdFxcDHzuPi4ET\nsS0bUkl9bHB9Bva7p2PdQHNJ7KAajrDum4kSSDYDB4Bqec5Xw75NBPI7h0bQan7PJapw6iJHJ+Bt\nbBZG3j7kRBNOPXQB7sQmGmQDTwFlnZ8vi04xYyKcuvgNGxf4we/cD8771Il0AWMonLq4EsuWcSs2\no/EjYCBwCtbVk0rCum8mSiD5G5tlc1qe892xzbEC+RQ4mdyDRt2xRY15c3QlknDqAmwa4NvYjfSx\n6BQtpsKphxbYjJycx2hseujxJPbEg3Dq4mPgMOzbeI76WLdQqv3/kYYNrPvLOU6Ue2SkJOt98x9K\nr+IKtS7Ssal/D2DfLqo7j0QfWA21HvLKIHlmbYVaF2lY12YWtltpK2ztQEFfRhJFqHXRCWvFjMLW\njrTGpv9uAA6PSYmjpyz237cldg8Y5fycivfNfyi9iiuUupiB/Y9yMM8jb30lolD/JvxlkDzrSCD0\nuqiOjQNsxwZUZ5H4Xy5yhFoX52GBdQdWF69hkw8SXTru/+/+94CnnedT7b4pIiIiIiIiIiIiIiIi\nIiIiIiIiIiIiIiIiIkkuC2/SumQQ/yvfM7GMvQWpiy1Cax3C+64niTYzEpHkVhX4D+4q5N+xlA3d\n/K6phKV6iLUM4j+QlMfSWuTIAibluaYYcBSWRytYVcidHuQgli5D5BCHeV0ASXkvA6WBS7Dss9Ww\nrKv+O9Nt86BciSKYQHcQ248jFP8LcC7VUqqLSAKohN3kuhZyXRa5v2VXA+Zhm1OtxzZm+i+W2TjH\nQWzHuxex3e/WYRta+bsfWOX3Pg+QO+tpBoXfqK/EtnHdA/yJJfvz/+Y/BFjpPL8aGE7uG3Iw5RyN\nJRDciyXUm+n3XCZu11Ymh+ZTq0Purq1iWMr0a/N8RmPnmpbO8Qbcrq0NHJqj7Rjn5xPyvM/lTj3o\nS6qIxMRhWMLAieS+gee1iNxjJPOBr4F2WAr4hc77jPa75iB2w7wIS49+H5YRtrbfNSOx/SbqAGdi\nabLH+D2fQcGBpA22l0l/533/BVyPG0guB37FuoSOwXam/A24JoRy9sWyr54J1MJu3Ff7vX4GFlTB\nurg+wfZZOcp5FOPQMZIHsHTh/u7GgnEO/zGSI53XX+K8ZxXn/Hzg8Tzv8ynwMCIiMdQH60bZg6Uw\nH4/t0ufPP5A0wW5q/tfUwjZpyhtIxvodF8fSaF9UQFmuAtb6HWdQcCDpg3W7lcvn+Y0c2roYDqwI\noZw3Yq2m/L7hZ5J7sD1v0IVDA8m/nGP/vUjWArf5HecdbA80RtIX29M850vAsc51zfIpqySpVNu0\nReLPK9he6r2Ad4COwGfAiHyub4rdrPy3Cf4Z++af17d+Px/AulyO8jt3HrbB029YwHiE3C2WwizA\nWjHrgWexLracoFIVC3DTnPfOeYwj9w28sHLOxcaQ1mMtjfOAkiGUMZBvge9wg1w7p0zPhfg+87CN\npHICzCXAUqwrT1KIAonEg31Y99Q92J7y04G7KHo/e3aeYx/u33x7bPOed7Aup5ZYV1coN+md2Lf8\nC7DWxwis9VDD73OuJPeujM2dR7Dl/BlrhV2Jdd89jO0AWCaEcgbyLG4gGYBtL/tTiO+RDTyDBZDi\n2J7v04tYLklACiQSj77HgkjpAM+twv5u2/idq4W1akJxErZpz1jsxrwO6wIK1QGsO+l2rMuoLHAW\ntjnSr0BDbHA67yMU+7Btkm8E2mKBqGM+1/5NcAH4Bads7bBA+Gwh12cTePrwU0AXbNynHDA7iM+W\nJKOZFeKlKthspelYV8sOLEDcgrVQdjrXpeHOdFoNvAs8ge2Ctw8bV9mDfZMviP9sqdXA0dhYxGfA\n6cCFIZb/LOxm/CE2VtAFW9fxvfP8ndhss21Yy6cE1oKpic0YC6acGdgN/HOsPvphwWLtoS8DbIbV\nidjg/i4CT+MFa+ksBqZig/QvFlCenPfthrVc9gFbnfNrsO7BB7HgtDPQi0VEoqUk1iL4HLsR78Ju\nTA9hU4Nz5B1Azpn+uwe7wQ3G1qDc7HdNoMHhvAPI92HrK3YAL2GD7Qf8ns+g4K14TwI+ADZjU4i/\ndcri70KsxbPH+R0/xFoAwZazNzYJYSt2k14K9PC71n/WFtie40uwujyAO/33AIeubB/ifP5LAX63\nvHXVE/tv8zeHtqgGOe/TCRGRBHUk9i35XK8LkqJuxbocRUQSRhfsm3p9bND8fayr5vCCXiQRVxYb\nr/kF62YUEUkYp2HdSDuxrqnXOXRKrURfJrbafg6auCMiIiIiIiIiIiIiIiIiIiIiIiIiIiKSCP4f\nkNsJwTerhcUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x884f550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "report.roc()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calibrating results $p(\\text{vrt same sign}|B)$ and combining them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "models = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from utils import get_result_with_bootstrap_for_given_part"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBEAAAFCCAYAAACn2UWVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkiRJkiRJkiRJkiRJUrX8/7s7xgpwcrAaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xe389d50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC for tagged: 0.594531402965 AUC with untag: 0.545941336892\n",
      "mean AUC after calibration: 0.594617024804 1.39432754719e-06\n"
     ]
    }
   ],
   "source": [
    "models.append(get_result_with_bootstrap_for_given_part(tagging_efficiency, tagging_efficiency_delta, tt_folding, \n",
    "                                                      [data_sw_passed, data_sw_not_passed], \n",
    "                                                      logistic=True, name=\"tt-log\",\n",
    "                                                      sign_part_column='signVtx', part_name='vertex'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAFCCAYAAABIGy6rAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZIkSZIkSZKk2vD/A7+GjsFaNfpLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x8839ed0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC for tagged: 0.594426504272 AUC with untag: 0.546503619328\n",
      "mean AUC after calibration: 0.594335879589 8.77876636993e-07\n"
     ]
    }
   ],
   "source": [
    "models.append(get_result_with_bootstrap_for_given_part(tagging_efficiency, tagging_efficiency_delta, tt_folding, \n",
    "                                                      [data_sw_passed, data_sw_not_passed], \n",
    "                                                      logistic=False, name=\"tt-iso\",\n",
    "                                                      sign_part_column='signVtx', part_name='vertex'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABA4AAAFCCAYAAABxfJ/wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x3dccbd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC for tagged: 0.594876938573 AUC with untag: 0.545332191596\n",
      "mean AUC after calibration: 0.594915538591 1.91431386544e-06\n"
     ]
    }
   ],
   "source": [
    "models.append(get_result_with_bootstrap_for_given_part(tagging_efficiency, tagging_efficiency_delta, tt_folding_rf, \n",
    "                                                      [data_sw_passed_selected,\n",
    "                                                       data_sw_passed_not_selected,\n",
    "                                                       data_sw_not_passed],\n",
    "                                                      logistic=True, name=\"rf-tt-log\",\n",
    "                                                      sign_part_column='signVtx', part_name='vertex'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAsAAAFCCAYAAACXVVS0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1078e150>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "models.append(get_result_with_bootstrap_for_given_part(tagging_efficiency,\n",
    "                                                       tagging_efficiency_delta, tt_folding_rf, \n",
    "                                                      [data_sw_passed_selected,\n",
    "                                                       data_sw_passed_not_selected, \n",
    "                                                       data_sw_not_passed], \n",
    "                                                      logistic=False, name=\"rf-tt-iso\",\n",
    "                                                      sign_part_column='signVtx', part_name='vertex'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparison of different models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>name</th>\n",
       "      <th>$\\epsilon_{tag}, \\%$</th>\n",
       "      <th>$\\Delta \\epsilon_{tag}, \\%$</th>\n",
       "      <th>$D^2$</th>\n",
       "      <th>$\\Delta D^2$</th>\n",
       "      <th>$\\epsilon, \\%$</th>\n",
       "      <th>$\\Delta \\epsilon, \\%$</th>\n",
       "      <th>AUC, with untag</th>\n",
       "      <th>$\\Delta$ AUC, with untag</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>    tt-log</td>\n",
       "      <td> 39.742696</td>\n",
       "      <td> 0.073143</td>\n",
       "      <td> 0.027520</td>\n",
       "      <td> 0.000569</td>\n",
       "      <td> 1.093730</td>\n",
       "      <td> 0.022685</td>\n",
       "      <td> 54.594134</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>    tt-iso</td>\n",
       "      <td> 39.742696</td>\n",
       "      <td> 0.073143</td>\n",
       "      <td> 0.031223</td>\n",
       "      <td> 0.003030</td>\n",
       "      <td> 1.240869</td>\n",
       "      <td> 0.120423</td>\n",
       "      <td> 54.650362</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td> rf-tt-log</td>\n",
       "      <td> 39.742696</td>\n",
       "      <td> 0.073143</td>\n",
       "      <td> 0.027696</td>\n",
       "      <td> 0.000673</td>\n",
       "      <td> 1.100699</td>\n",
       "      <td> 0.026814</td>\n",
       "      <td> 54.533219</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td> rf-tt-iso</td>\n",
       "      <td> 39.742696</td>\n",
       "      <td> 0.073143</td>\n",
       "      <td> 0.028914</td>\n",
       "      <td> 0.003222</td>\n",
       "      <td> 1.149137</td>\n",
       "      <td> 0.128051</td>\n",
       "      <td> 54.678429</td>\n",
       "      <td> 0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        name  $\\epsilon_{tag}, \\%$  $\\Delta \\epsilon_{tag}, \\%$     $D^2$  \\\n",
       "0     tt-log             39.742696                     0.073143  0.027520   \n",
       "0     tt-iso             39.742696                     0.073143  0.031223   \n",
       "0  rf-tt-log             39.742696                     0.073143  0.027696   \n",
       "0  rf-tt-iso             39.742696                     0.073143  0.028914   \n",
       "\n",
       "   $\\Delta D^2$  $\\epsilon, \\%$  $\\Delta \\epsilon, \\%$  AUC, with untag  \\\n",
       "0      0.000569        1.093730               0.022685        54.594134   \n",
       "0      0.003030        1.240869               0.120423        54.650362   \n",
       "0      0.000673        1.100699               0.026814        54.533219   \n",
       "0      0.003222        1.149137               0.128051        54.678429   \n",
       "\n",
       "   $\\Delta$ AUC, with untag  \n",
       "0                         0  \n",
       "0                         0  \n",
       "0                         0  \n",
       "0                         0  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pandas.concat(models)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Implementing the best vertex model\n",
    "\n",
    "and saving its predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from utils import prepare_B_data_for_given_part"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KFold prediction using folds column\n",
      "KFold prediction using folds column\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAFCCAYAAABIGy6rAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZIkSZIkSZKk2vD/A7+GjsFaNfpLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13f00550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC for tagged: 0.594426504272 AUC with untag: 0.546503619328\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEACAYAAABI5zaHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIhIOPh/7/SOJ2tIaHIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13f00d10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Bdata_prepared = prepare_B_data_for_given_part(tt_folding, [data_sw_passed, data_sw_not_passed], logistic=False,\n",
    "                                               sign_part_column='signVtx', part_name='vertex')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Bdata_prepared.to_csv('models/Bdata_vertex.csv', header=True, index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}