data_balancing_ds1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Balancing "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import libs.config\n",
    "from pandas import read_csv\n",
    "from matplotlib.pyplot import figure, savefig, show\n",
    "from libs.ds_charts import bar_chart"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since we have different scalling versions that lead to different training files we need to spit the balances into two balances: one takes minmax_train and other takes zscore_train. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename_minmax = 'data/dataset_1/nyc_collisions_scaled_minmax_train.csv'\n",
    "file_minmax = \"nyc_collisions_scaled_minmax_train\"\n",
    "\n",
    "filename_zscore = 'data/dataset_1/nyc_collisions_scaled_zscore_train.csv'\n",
    "file_zscore = \"nyc_collisions_scaled_zscore_train\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## For minmax"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 97\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 0.0 : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "original = read_csv(filename_minmax, sep=',', decimal='.')\n",
    "# drop this because later, knn does not handle not float values.\n",
    "#original.drop(['CRASH_DATE','CRASH_TIME'],axis=1,inplace=True)\n",
    "class_var = 'PERSON_INJURY'\n",
    "target_count = original[class_var].value_counts()\n",
    "positive_class = target_count.idxmin()\n",
    "negative_class = target_count.idxmax()\n",
    "#ind_positive_class = target_count.index.get_loc(positive_class)\n",
    "print('Minority class=', positive_class, ':', target_count[positive_class])\n",
    "print('Majority class=', negative_class, ':', target_count[negative_class])\n",
    "print('Proportion:', round(target_count[positive_class] / target_count[negative_class], 2), ': 1')\n",
    "values = {'Original': [target_count[positive_class], target_count[negative_class]]}\n",
    "\n",
    "figure()\n",
    "bar_chart(target_count.index, target_count.values, title='Class balance')\n",
    "savefig(f'images/{file_minmax}_balance.png')\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see the data is not balanced..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So we spilt the data into two datasets:one for each class. Then we can sample the required one and join the other, as we did on the other preparation techniques. In the end, we can write the dataset into a new datafile to explore later."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_positives = original[original[class_var] == positive_class]\n",
    "df_negatives = original[original[class_var] == negative_class]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then we can use two strategies:\n",
    "    - undersampling (for huge dataset and a considerable number fof positive records): keeps the positive records and samples the negative ones to balance the final distribution (negative side is that this loses data)\n",
    "    - oversampling (small number of positive records): in order to create larger support the training step - this can be done with smote"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 97\n",
      "Majority class= Injured : 97\n",
      "Proportion: 1.0 : 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import concat, DataFrame\n",
    "\n",
    "df_neg_sample = DataFrame(df_negatives.sample(len(df_positives)))\n",
    "df_under = concat([df_positives, df_neg_sample], axis=0)\n",
    "df_under.to_csv(f'data/dataset_1/{file_minmax}_under.csv', index=False)\n",
    "values['UnderSample'] = [len(df_positives), len(df_neg_sample)]\n",
    "print('Minority class=', positive_class, ':', len(df_positives))\n",
    "print('Majority class=', negative_class, ':', len(df_neg_sample))\n",
    "print('Proportion:', round(len(df_positives) / len(df_neg_sample), 2), ': 1')\n",
    "# here we split by the minory getting 97 of each"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 27220\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 1.0 : 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import concat, DataFrame\n",
    "\n",
    "df_pos_sample = DataFrame(df_positives.sample(len(df_negatives), replace=True))\n",
    "df_over = concat([df_pos_sample, df_negatives], axis=0)\n",
    "df_over.to_csv(f'data/dataset_1/{file_minmax}_over.csv', index=False)\n",
    "values['OverSample'] = [len(df_pos_sample), len(df_negatives)]\n",
    "print('Minority class=', positive_class, ':', len(df_pos_sample))\n",
    "print('Majority class=', negative_class, ':', len(df_negatives))\n",
    "print('Proportion:', round(len(df_pos_sample) / len(df_negatives), 2), ': 1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import concat, DataFrame\n",
    "\n",
    "df_pos_sample = DataFrame(df_positives.sample(len(df_negatives), replace=True))\n",
    "df_over = concat([df_pos_sample, df_negatives], axis=0)\n",
    "df_over.to_csv(f'data/dataset_1/{file_minmax}_over.csv', index=False)\n",
    "values['OverSample'] = [len(df_pos_sample), len(df_negatives)]\n",
    "print('Minority class=', positive_class, ':', len(df_pos_sample))\n",
    "print('Majority class=', negative_class, ':', len(df_negatives))\n",
    "print('Proportion:', round(len(df_pos_sample) / len(df_negatives), 2), ': 1')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now the smote method splits the original data into two: one with just one variable - the class variable, call it y, and another with all the other variables, call it X. Then the smote generates the positive records, don't needing to join the positive and negative ones. \n",
    "So, What we have to do is just rejoin the data (smote_x) with the corresponding class (smote_y), already updated.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 27220\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import Series\n",
    "from imblearn.over_sampling import SMOTE\n",
    "RANDOM_STATE = 42\n",
    "\n",
    "smote = SMOTE(sampling_strategy='minority', random_state=RANDOM_STATE)\n",
    "y = original.pop(class_var).values\n",
    "X = original.values\n",
    "smote_X, smote_y = smote.fit_resample(X, y)\n",
    "df_smote = concat([DataFrame(smote_X), DataFrame(smote_y)], axis=1)\n",
    "df_smote.columns = list(original.columns) + [class_var]\n",
    "df_smote.to_csv(f'data/dataset_1/{file_minmax}_smote.csv', index=False)\n",
    "\n",
    "smote_target_count = Series(smote_y).value_counts()\n",
    "values['SMOTE'] = [smote_target_count[positive_class], smote_target_count[negative_class]]\n",
    "print('Minority class=', positive_class, ':', smote_target_count[positive_class])\n",
    "print('Majority class=', negative_class, ':', smote_target_count[negative_class])\n",
    "print('Proportion:', round(smote_target_count[positive_class] / smote_target_count[negative_class]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Charts!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjIAAAGxCAYAAAB4AFyyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAABcSAAAXEgFnn9JSAABPx0lEQVR4nO3deXwU9f3H8dfu5r4I4Qq3IKeAIALirVURUTx/WrXaqq2/revaigce9T6q1ir96bh2W61KrVUrVotXFW9RORSRWw6BcIQrCeRONju/P2YSlk0gBwu7A+/n47GP3Z3vZ77z3clO8sl3vvMdl2maiIiIiDiRO94NEBEREWkrJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcKyneDRCRg1f2pOlmC0M/LZ0y8aR92Za2yp40/QagqHTKxOfj3RaRg5ESGRGJp8uj3p8PnAfcBGyKWL6JxHUDsAJ4Ps7tEDkoKZERkbgpnTLxxcj32ZOm98NKZN4snTJxRSy2kT1pembplInlsahLRBKPEhkRSWjZk6YPAiYBJwE9gVpgFnBX6ZSJX0fFrgbWYfWS/AEYDXwAnGuX/xb4DdAdWAU8AvQB7i6dMtEVVdcQ4G7gZCAH+BH4KzCldMrEsB1Tf2qse8TrNaVTJh4Si88uIs1TIiMiie4krITkFaAA6Az8Cvg4e9L0I0unTFwcFd8DeA/4h/2oAsieNP124EHgS+AJIA94DFgbvcHsSdPHAjOwkqLHgWKshOaPwKGAzw693K5rk103QNlefl4RaQUlMiKS6KaWTpn458gF2ZOmB4HFwG8Bb1R8b+AXpVMmTo2I7wDchZXEnFQ6ZWKtvfwV4Puoul3AM8BS4NjSKROr7aI/Z0+aXgDckD1p+hOlUyYuLZ0y8cXsSdMfBjZFnyYTkf1DiYyIJLTSKRMr6l9nT5qeAaTbb2dj9dREKwaik4rTgFTgyfokxq57cfak6f8FJkTEDgOGANcD2dmTpmdHlL0L3AicgpXoiEicKZERkYSWPWl6DvAAcCGQH1X8YxOr/Fg/hiXCIfbzD03E/8Cuicwg+/lP9qMpnXezXET2MyUyIpLoXgZOB54CZmL1uISB27DGq0Sr3Mvt1Q/6vRf4YjcxTSVQIhIHSmREJGFlT5qeC5wBPF86ZeJvosrua0VVq+3ngcC3UWUDot7XX/ZdVTpl4owW1N3SSf1EZB/QLQpEJJHVYSUKu/yuyp40/XhgbCvqmQFUA9dlT5re8A9c9qTph2H19kSahzX+5frsSdO7RFeUPWl6dvak6WkRi8qB9q1oi4jEkHpkRCRhlU6ZWJo9afpHwGXZk6aXAd8Bg4FfAouA7D2sHlnP1uxJ0x8E7gM+ta9W6oB1GfV8YGREbDh70vRfYCU/i7MnTX8WWI6VrAzFmn14BDt7buYCl2ZPmn431nibstIpE6fvzecWkZZTj4yIJLpLgZewBvv+Casn5nzgm9ZUUjpl4v1YE+V1BR4F/gfryqSPiRpXUzpl4mzgSGA68DOs8Tk3YE2edy+wPiL8Nqx5a2602/lka9olInvHZZo6vSsiB6/sSdP/AwwsnTJxYLzbIiKtpx4ZETkoZE+ant7EsiFYg4k/3P8tEpFY0BgZETlYnJk9afotwBvAZqyrlbxYtxR4JI7tEpG9oERGRA4WS4AtwHVY91kqxbqh5B2lUyauiWfDRKTtNEZGREREHEtjZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpauW9pIRCLqxpjqPtM3v84bj0R4REZGDiRKZvdcBa06KSJ2xLvMUERGRfUinlkRERMSx1COzD/zyyp9H99CIiIjInrnaspJ6ZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpcG++5lpmuhGnbK3XC4XLlebxsWJiBxQlMjsJ1VVVWzYsIGamholMrLXXC4XKSkpdOvWjbS0tHg3R0QkbnRqaT+oqqpi7dq1VFdXK4mRmDBNk+rqatauXUtVVVW8myMiEjfqkdkPNmzYQF1dHWlpaXTv3p2kJO32RBE2TdYXV1JaHSIUNkn2uOiUlUrHrFRqQnUs3VTWKD4nNZm+nTKprQuzoaSSspo66sImqUluuuSkkZueDEBVbR0bt1dRXhMibEJasptu7dLJSt3586+sraOguILK2jqSPVZ5/fpNqayto6CogsqaWigvonx7JR98MI8HZhbTtV0aFx/Zg/eXbOa7dSXUhExSktwkua1TUFWhMOMGd+bVX41hS2k1t76xiC9WbqO0KkSfjhncPn4gZw7NB2D55jLuemsJs1cXU11bx6D8bO6feBhH981raMvSTaX4X57P9+u30y03nQfPPqxh/aY0F9/a+uKlOlTHjdMW8skPW9hWXkPXdmlc/5N+/PyoXhQUVzD64U92idd+jw0n7/eu7dI4JC+DlVvLG9quYzWG6sds6NG2x5NP/bnTk0/92Yx8VFRUmPXC4bC5ZMkSc/HixWZ1dbUpiSVUFzbXl1SYVTUhMxwOm2VVteb8ghJze0VNo9i6cNicv67E3FZm/RyrakJm4fZKs7q2zgyHw2ZJRY05r6DErKgJmaZpmmVVteaW0iqzJmSVbymtMr8rKDFrQnUN9S3csN1cX1Jh1tVZ639XUGJW2us3tf3I+M0lpeasb783Z337vVlXV2fOXLnVzJn0H/OqqXPNypqQ+c7CQrPL5LfNFZvLzOraOrPX7941X5lbYJqmaa7aUmb+6aMV5rpiq663F2w0O09+21yycYdpmqY5Z3WR+ezM1ebm0iozVBc2//blarPbre+YW0qrTNM0zZpQnXn4AzPM+99Z0mhbTWkuvrX1xVNZVa15/ztLzJVbysxwOGzO+rHI7HHbu+aMJZsaxWq/x46T9/vr89ab7W+Ybn68bLMZDod1rO5em/4O69TSPla/owH1xCQgj9tFt3bppCZ7cLlcZKYmkZWWRFlNqFHs9spagIYek9RkD11y0khJcuNyuWiXnkxakpvyamvdzNQkOmalkuyxyjtmpeJyWT01AOXVIerqTPJz0nC7rfUzUz0UV9Q22dbo+A7ZGaQkuSizt1cTCuNxueiVl0FasoczhnTh2EM78PLcdby1oBDThLMP7wpAn46Z/PbkQ+mem47b7WLC0Hz6d8pk9ppiAEb1bs9Vx/SmU1YqHreLK4/ujcftYtHGUgBmrtxGUXktt4wb0GhbTWkuvrX1xVNmahJ3nDGIvh0zcblcjDmkPcf368BXPxY1itV+jx0n7/fzRnTj5IGdmLmyCJfLpWM1xpTIiEQIh00qaupIT/Y0KttWVkP7jGTc7qavFqqtC1MVanpdgEr7FFSaXV5ZW0dasht3xNVHGckeKu1Ep9H6TcS7Xa6G+HkFJbhcMLx7u4byw7vnsHDjDqbOWstFI7s3bDvaltJqlm0uY2jXnCbLF23YQWlViEFdsgBYuGEHg/KzSPbs/BVSv62mNBff2voSSVVtHd+sLWly32m/7ztO3u86VmNLXQRxYprWH8x4yUjx6PLdKKZpsra4gtQkd6NxKtWhOkqrQ3TLzWpy3bBpsnpbBbnpKWSmNj6sQuEwP24rJz8nreEXQF3Y6hGK5HG7qAs3PSC8qXiwki/TNHn12/VkpiY1/CcH0C49ma2l1cxeU8z9Ewc3WW9NKMwVU7/hvOHdGNkrt1F5SWUtV0z9hptO7U+XHOsKqfKaOtpF7aN26cmUVTXuyWpJfGvrSxSmaeJ/ZT6HdsrcZb8DrC2q4OMftmi/7wNO3u86VmNPiUycVNTUkX/ru3HbfuHDZzT5B/dgZZomBcWVVNWG6dc5s1GSV1ReS3qyh4yUxvssbJr8uLUCtwt65aU3Kg+Fw6zcUk5WahL5OakNyz1uKzmJVGeaTSYru4sHcLtdTP73QooraunXKXOXHqMdVSG2VdQwvHs7hkX891evJhTmsufnkp7iwfjp8Ebl2ytrOe/PX3N03zxuHz+gYXlmiocdlbv+4tpRFSIrrenvVHPxra0vEZimyaTXFrB8cxn/ueboRj11L84u0H7fB5y83zNTPUx6bYGO1RhL3G+rHHCyJ03n6UtGcNmYni1ex/vSPNYWVfKu/5iYtuXzFVuZ8NRXLLzzFHq1T2ddcSUVNXX065SJi+heEZNt5TV0zk5t1FtimiZriioIm9CnQwamaSUj9erCJqu2lpOW7KZ7bhphE7DLU5M8VNVWE6oLNyRO9ae1muqViY4Ph03CYZOaUJi5a0q4/6zBTJq2gPUlFbRLS8blcvHt2mIKt1dzxxmHNIzdqVcTCnP1P+ZRUxfm+Z+PpLYuTG1EprSjqpaLn53DgM5Z/P7sw3bpQezXKZMlhaWUVNQ09DDNKyhhWLecRttpSXxkebt0q+3fr9/OiB6Nf6EnAtM0uWHaAuauKeY/14wlye3a5XOHwyZTZ63lNycdqv0eQ07e79+uLWZLWQ1riiq5Y/wAbn1zMcXl1aQkeRrKnXSs1oTqGtoe7++MyzSb7saWljECwU7A5shlv7zy56SnW/+Zh8Nhli1bBsDAgQNxu60vRnl1yHE9MqVVIR7/cDlvzN/I2qJKMlI8jO6dy6RT+nF8v47Nrr9pRxU5acmkpzR97rcp2ytrqQub5GWmtKqtzYlMZNxAWU0d/TtZPTHz12+P6bb2FdMMQ/EGlm0q5br3txLHM5UxteaBccxZU8IvXviGmTedyKGdMuPdpEZueG0BX/1YxFu+o0lLcsf1WI4V7XdpjRtP6cdt4wfwyQ9bY/mdadN4B/XISIvsqKpl3BMzqayt454zB3Nkr1xKKmv5+6wCzgp8xZM/Hc7Pj+rV5Lo1oTAp9hwrrRV9LjbWakJ1lFbX4QYWbSxFaX38DbxnBt1z03nmspEJ+cd0bVEFf525mtQkN0Pum3HAfGe036U1PluxladuX0X3dvH/ziiRkRa5/51lrNxazje3nkyvvAwAegGPnt+O2rowN05bwGmDOlNTF2bo/R/yl0tHMO27DXy+YhuXjenJYxcMa3RqafW2Cn776nxmriqiY1YKN5zSj39/t5FeeekELz0CaHxq6ffvLeMfswt45Lyh3PXWYtaXVDGyZy5P/vRw+nWyBuIWV9Qw+fWFzFxVxJayarq1S+PCkd2ZfNoAUpJ2vVAvJcnDyM7ZDe/rwmZDj8zgzG24E/jXZThssmp7iMOytjH36BcpqvZw0pxLADj3uG0keRK37dFCdS7e+KIDAKsfOD2hx2/1ysugdMrEhvfx7l2NFSfv91ljXyTDk9gDlOttq0k9II7T6dccnTDfl8RohSQ00zR55Zt1XDSyR0MSE+mmU/vz7JdreGP+BibYszve9dYS7j5zMI+eN3S3dV76tzkke1y86z+GtCQ397y9lO/Xb29ywGykLWXVPPvlap69bCRJbhe//ud3+P45n/d/cyxg9QANzM/Gd2Jf8jJT+K6ghOtfW4DH5eK28QNb/LndmOxm3G1icFn9sG5M0j0h0iPO2CV5TJJafgYvATjnl7kkngxPyDGJTIVn559dHaexoURGmrW1rIbiiloG5zd96XGP9unkpCWxfHN5w7JfjO21x0G9H/+wlQUbdjD31pMY2MXqEfnrz45g0L0fNNueqlCY4KVH0DnbugLoNycfytX/mEdVbR1p9iR1N53avyG+d14Gq7ZW8MKsta1KZEREJPEpkZF9YlSv9nssX1pYSm56ckMSA5CXmUL/zk0nS5Hyc9IakhiArjlpmKbVU9OzfQbhsMkTn6zktW/Xs66kiqraOkJhk7AGtouIHHCUyEizOmSm0D4jmcWFpU2WryuuZEdViP6ddw72ykxtvr+0rfPxJUed76mvp/6K5Sc/XcUj7//Ag2cP4cheuWSmenjt2w08+N6ytm1QREQSlm5RIM1yu11cOLI7//p2PWuLKhqV/3HGclKT3Jw7vFuL6xyUn01xRS0/bN55d+niihpWbCnfw1ot88WKbfxkYCeuOqY3w3u0o1+nLFY30W4REXE+JTLSIndOGMQheRlMfPor3vhuAwXFFSxYv52bX1/Ic1+t4bELhtG1Xcsvrz55QEeGdcvB+495zF1TzIL127n6H/NIcrv2+tYJA7pkMXNlEZ/8sIUVW8qY8uEK3llYuFd1iohIYtKpJWmR3PRkPrz+OB6bsYJ73l7K2uIKMpI9jD6kPdOvOZoT+jc/IV4kl8vFS1eN5jevzme88WXD5ddby2pIS9q7/PqWcf1ZX1LJz56biwuYMDSfW8cN4JY3Fu1VvSIikniUyMRJRoqHwofPiOv2WysnLZl7zxrMvWc1fUMzsK4QipzrIVL08kM6ZPCfa47eWV4V4t63l/I/R+w8RVU/n0y928cP5PaoK4+O79dxl7pz0pJ5/udHNtq+78S+u11HREScSYlMnLhcroSZTChe3l20CRfWeJktZdU8/N8fME244Iju8W6aiIg4xMH9l1Tiqqq2jgffW8baogoyUpIY0aMd7//m2FaNtRERkYObEhmJm/NGdOO8ES2/0klERCSarloSERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYG+8aJaZpU1NTFbfsZKZ69nkFXREQk3pTIxElFTR35t74bt+0XPnzGQT+PjYiIOJ9OLUnceV+axxnGl/Fuxn418YU1XPvmhng3Q0TE8fQveQKY+7OvyEja96eZKkIeRv3j6OYDo5xhfEmvvPRGtwsAyJ40nacvGcFlY3rGool7zTRN/vz5j0ydVcCPW8txu1x0y01jbJ88jJ8Oj3fzREQkxpTIJICMpDoyksPxboajherCeNwu/vDBcp74eCWPnDeUo/vkEQqHWbapjHcWbYp3E0VEZB9QIiMxkz1pOn84bwjfri3hrYWF5KQlc/Vxh3DTqf0bYorKa5j02gLeW7yJzBQPVxzdG9NsXNdr365nykcrWLapjM7ZqZw9LJ87JwxqGNfjfWkea4sqOX9EN574eCVriytYed84/vP9Ri4b03OXHqKBXbI5+/CuDe+LK2qY/PpCZq4qYktZNd3apXHhyO5MPm0AHrc1AHpeQQkn/msJL1/cgz/N3MZ3G6vo3T6ZKWd2pVduMje+XcgXq8vpkp3Eg+O6cPqA7Ib68+5bwgPjOjN3XSUfLC8jPdnNFUe257aTOu5xgPVfZhfx7Jxi1pbU0r1dEj89vB3XH9uRZI8GZYuI7E5cExkjELwZOB8YBLiAhcADfp/3vYiYk4CPm1j9Sr/P+3xEXFfg/4Dx9qL3gOv8Pu+miJhk4EHgciAX+Ba43u/zzolq1xXAbcAhwGrgQb/PO7Wtn/Ng8ugHy7lzwiBuGz+Q/y7exOR/L2JUr1xOGtAJgGtfmc/C9Tt4+arR5Oek8diHy3ln4SaG92jXUMeLswu47Y1FPHLeEMb2yWPD9ipumraATa/M57mIu1p/t66EzFQPL145ihSPm5y0ZPJz0vhyVRHriivp0T69yTbWhMIMzM/Gd2Jf8jJT+K6ghOtfW4DH5WLyuAG7xN770RbuO7UzvXKT+d37m/jltPX075DClaPa88C4Lvxp5laufn0DC67vR7u0nXcUf/SzrdxwXEduP7kTX6+tZPK7hXTK9HD1mLwm2/TwJ1t4aX4Jvz+9C8O6pPHD1hpueHsjlbUmd53Suc0/DxGRA128B/v+BPgbcDJwFDALeMsIBI9tInYM0DXi8Up9gREIuoG3gEOBcfajHzDdLqv3KPAr4NfAaGAFMMMIBLtH1HUu8CzwF2C4/fycEQhObOmHCofDuzxMu8vBNM1dHvEU3ZY9PaDp9jd8hojX5wzvyhVje9GnQwa/Pr4PfTtm8PGyLZimyYrNZby1oJA/XjCUkwZ0ZFB+Fk/9dDhZqR5gZ50PvbeMOycM5JJRPejTIYNj++bxx/OH8tq8DWwtrd5l3z3zsyM4vHsOg/KzSPa4eOicwwjVhRl83wyGP/ghV079hhe+WkNFdaih/s7Zqdx4Sj9G9GhHr/bpnH14V35zUl9e+XZ9w2etd9PxHTmlXxb9O6Zy4/Ed2Vga4if9sjjnsBwO7ZDC7Sd3oqwmzNx1lbusd3LfTK47pgP9OqRy2RG5/HJUe574cluTP4uK2jBPfrmNxyZ05axBOfRun8Jp/bP43cmdeGZOcdy/K/tb9PHjhMeBIN778GDd706VSD/PuPbI+H3eM6IW3WgEgmdg9dLMjCrb4vd5C3dT1anASGCo3+ddBGAEgpdj9fD8BCtZycFKYG7w+7xv2jFXAacD1wB32HVNBqb5fd7H7PdLjUBwLHALML0ln+vlV6c1nEJwuVwMH3YYqampVFZV4XZbeVVlHOeQAaisrsZthloUGw6Hqauro7KqqsnymtrahrJBnTN2ieuSlcKG7RVUVlXx/boiAIZ33TVmePccdlRZdWwtq2FtcSW/e3Mxd05f0hBT/8d80YYiRvfKpa6ujn4dM0lx1VFZtXNf9shJYoZ/DIsLy5i9poRvC3Zw65uLeHTGct759Wg6ZqUQDps8/cUa/v39JjaUVFEZClMXNgmbJlVRn/Hw/J134s7Psg6XYV1SGy3bUr7rvhzTM2OX92N7ZfDU10XsqK4jJ9WzS9nSzdVUhkyufG0dkSeR6kyoCplsKguRn53c1K4/IC1Zuoz05Hj/j9VylbUHxh9U7XdpjX3xfRk2dEib1kuoMTJGIOgBsoHyJoo/MQLBDKxelCAw1e/z1v+reixQUJ/EAPh93kVGILgOOA6YARwJpGKdcqqPqTMCwQ/sGIxAMAWrp+aZqG2/BzxtBILJfp+3du8/qbOkJLnZUdU46SmptHZFatLOL3Oye9fxHC4XTY6B2Z2wHXzvhAGc2L/xaZj8nJ1JREZK0weRy+ViSNdshnTN5sqxcGNRH459/Cuen7WOm07py59nruWxj37kngn9OaJHDpkpSfx7fiF/+HBVo7qSPZH11i9zRSxz2e1u+WeMVr/uM+d3Z0Cn1EblHTMT6jAVEUkoifYb8g6sROYvEcs2AtcCc4EwcCbwV6xTR3faMV2BpnprCu0yIp6j4wqxTlsBdMTaJ03FJNvlG5v7EBdfdAHp6db4jHA4zA8//IDL5SI9La2hRybs2pkYVIQ8TdYTa5HbSU9NJb2FE+INzm/HWwsLSU5OIcmzM3n44scdAAzt0Z70NKvnIiU5ueE1gNvtxuPxkJ6WxuE9rMRk/sYKxg22xn3UhMLMX7+DQztlkp6WRu+0NHrkprG6uJprujc9ngTA4/Hgdrt32dbuDMxPJSPFQ3FVHelpacxas51TBnbCe0K/hhjj87UApKWlATUt2i97Mruggl8ftbP9swoq6JaT1Kg3BmBQ51TSklz8WFzL+IHZjcoPNoMHDXTUZI3l1SGsoXTOpv0urZFI35fEaAVgBII+rNM6Z/t93nX1y/0+7zJgWUToXLvnZJIRCN6XiD0kbre7IWEBdjnNFPm6XlvmdtlbkW1pzq+OO4QXZq3F+8/5XHtiXzpkprBwww5ue2MRJ/XvyPAeuZEVR9W78/P265zFhKFdmPzvhaQlD6dLdiqPf7iCsuo6YOd6d00YhO/l+XTISmHisHySPG5+2FTGe4s3Rc0F0/gzXPq3OYzq3Z6j++TRPTeNLWU1PP3ZKkqrQ5w9rCsul4sBXbL5x+wCPl2+lR7t05n+fWHE5dmxuULo41XlPPXVNsYPyGZWQQXPzine7aDdrBQ3k47rwEOfbMHjtsbXhMKwZEs1322o5L7TusSkTU4RffwkOie1dU+036U1Eun7khCJjBEI3gTcA0z0+7wftWCVL4HbgU7ABqxekvFNxHVhZw9K/XM+sGo3MVuBkB0TXU+tXX7QGdA5i0+uP5773l3Kz56bQ1F5Db3yMrjoyO7ccEq/5iuIELh4BJP+9T0XPTOb9GQPvxjbi7OG5VNQvHOw7CWje5KdlszjH67gsRnL8bhd9OmQyVnDon8sjY07rDP/nreBpz9bRVF5Le0zkhnaLYfX//coTh5oXTl1y7j+rC+p5GfPzcUFTBiaz63jBnDLG4v2XHkr3HR8R2avq+ShT7aQnuzGNzaPq8e03238zSd0Ij87mb/OLuLuDzaTluzi0LwULh7ebrfriIhIAiQyRiB4H3A9cIbf5/28hauNBCrZmVjMBO4yAsHBfp93iV3vYUBP4As75hugGmtw79N2jBtroPDfAPw+b40RCM6xY/4Wsb3xwOxY9v5kpHgofDh6rPP+k5HSutNZg7tm88+rRu8xpnRK4wu73vUfs8v7DpkpTL1iVLPbO2tY/h4Tl6ZmGQa4Ymxvrhjbe49156Ql83zEZdz1fCf2pc4esHJEz1y23jmYyCE/vXJTKLprcKP1mlrWPt3DCxf22G0bpv+icRsvPyKXy4/I3WPbRURkV/GeR+ZPgBe4BFhuBIL1f7lq/D5vkR0zCVgLLMK6NvYMrN4Yw+/z1g9mmIE1J8xUIxD028ueAuYAHwH4fd4dRiD4Z+ABIxBcD6wEbgIysRMb2x+AafZ23wEmABcA58bys7tcroQ5vygiIuJU8T7B9VsgDfg31umd+sfrETFJwO+xEpXZWJPZ/QbrcmgA/D5vGDgLa+TXDPvxI9Z4m8hr9G5m5xwx3wADgdP8Pu/6iLreAH6JdUn2Avv5Sr/P26JLr0VERGT/ifc8Ms2OrPT7vI9iTWTXXNxG4MJmYmqxBhRPbibueeD55rYp0pSmTjWJiMi+Ee8eGREREZE2UyIjIiIijqVERkRERBxLiYyIiIg4lq7/jRPTNKmI440jM1I8LZ7ZV0REJFEpkYmTipo68m99N27bL3z4DM1jIyIijqdTSyL7wMQX1nDtmxvi3QwRkQOe/iVPALPGvkiGJ9R84F6qqEviqK8va/P6pVUhHv9wOW/M38jaokoyUjyM7p3LpFP6cXy/jjFs6e6ZpsmfP/+RqbMK+HFrOW6Xi265aYztkxd1Q0kRETkYKJFJABme0H5JZPbGjqpaxj0xk8raOu45czBH9sqlpLKWv88q4KzAVzz50+H8/Khe+2z7obowHreLP3ywnCc+Xskj5w3l6D55hMJhlm0qi7h7tYiIHEyUyEiL3P/OMlZuLeebW0+mV14GAL2AR89vR21dmBunLeDUgZ0Y9+RMfja6J7eNH9iwbl3YZOA9H3DNCX248dT+ALz27XqmfLSCZZvK6JydytnD8rlzwqCGcTvel+axtqiS80d044mPV7K2uIKV943jP99v5LIxPblsTM+G+gd2yebsw7s2vC+uqGHy6wuZuaqILWXVdGuXxoUjuzP5tAGkJFlnUz9fsZUJT33Fv341hsc/XMG8dSUckpfBny48nLLqEI99uILvC4rpkp3Eg+O6cPqA7Ib68+5bwgPjOjN3XSUfLC8jPdnNFUe257aTOu5xAPVfZhfx7Jxi1pbU0r1dEj89vB3XH9uRZI8GXYuItJUSGWmWaZq88s06LhrZoyGJiXTTqf159ss1vPn9Ri4Z1ZOXv1m3SyIzY+lmtpRVc8ko627QL84u4LY3FvHIeUMY2yePDduruGnaAja9Mp/nIu5K/d26EjJTPbx45ShSPG5y0pLJz0njy1VFrCuupEf79CbbWxMKMzA/G9+JfcnLTOG7ghKuf20BHpdrl3YB3P32Eh48+zB65WVw2xuLuHLqN3Rtl8Y5w7vyp/HteeLLrVz9+gYWXN+Pdmk77xj+6GdbueG4jtx+cie+XlvJ5HcL6ZTp4eoxeU226eFPtvDS/BJ+f3oXhnVJ44etNdzw9kYqa03uOqVzy38YIiKyCw32lWZtLauhuKKWwflZTZb3aJ9OTloSyzeXc8noHqzaWsGXq7Y1lP9z7jpOHtCJbrlW4vHQe8u4a8IgLh3dk74dMznu0A788YJhvDZvA1vLqnep+9nLRjK8RzsGd80mJcnNQ+cOIRQ2GXzfDEY8+BFX/f1bXvh6LZURl7J3yUnjplP7c0TPXHrnZXDO8G785qRDeeXb9US75bQBnDqoMwM6Z3Hzaf3ZuKOaMYfkcdKAThzaIYXbT+5EWU2Yuesqd1nv5L6ZXHdMB/p1SOWyI3L55aj2PPHltkb1A1TUhnnyy208NqErZw3KoXf7FE7rn8XvTu7EM3OKMU2zZT8IERFpRD0yElN9O2ZydJ88/jlnHcf07cCOqlreXlhI4OIRAGwpq2ZtcSW3/2cRd05f3LBe/Z/ylVvL6ZiVCsCAzlm0S0/epf4BnbP48qYTWLSxlK9WFTFnTTG3vbGIx2Ys58Prj6NTVirhsMkTn6zktW/Xs66kiqraOkJhk3ATCcPhPXIaXufnpAHQr3PmzmVZ1iGypXzXMUxjeu7aMzW2VwZPfV3Ejuo6clI9u5Qt3VxNZcjkytfWEXkSqc6EqpDJprIQ+dm7fk4REWkZJTLSrA6ZKbTPSGZxYWmT5euKK9lRFaK/nQBcOqYHd7y5mEfPH8rr320g1ePmrKH5ANTnEg+dM4STB3RqVFe3dmkNrzNSmv56ulwuhnbLYWi3HK4+7hB+3FrOyIc+5pkvVnPb+IE8+ekqHnn/Bx48ewhH9solM9XDa99u4MH3ljWqK9njjqjXek5yuyKWWa/De9FpUr/uM+d3Z0Cn1EblHTN1GIqItJV+gyaAirr982No63bcbhcXjuzO1FlruXXcgEbjZP44YzmpSW7OHd4NgPNHdGPy6wt5a0Eh/5yzjvNGdCM9xeql6JydSo/cNJZvLuNXxx6yV5+nXu+8DDJSPGwpqwHgixXb+MnATlx1TO+GmNVFFTHZVr3ZBRX8+qid42FmFVTQLSepUW8MwKDOqaQlufixuJbxA7MblYuISNspkUkAezO3y/5y54RBfLZ8KxOf/op7zxzMkb1zKamoZeqsAp77ag1PXDScrnZvSk5aMmcN68rjH65g4cYd3HPm4F3qumvCIHwvzycvM4WJw/JJ8rj5YVMZ7y3e1OxcMJf+bQ6jerfn6D55dM9NY0tZDU9/torS6hATh1m9PgO6ZPGP2QV88sMWerRPZ/r3hbyzsDCm++PjVeU89dU2xg/IZlZBBc/OKd7toN2sFDeTjuvAQ59sweO2xteEwrBkSzXfbajkvtO6xLRtIiIHEyUy0iK56cl8eP1xPDZjBfe8vZS1xRVkJHsYfUh7pl9zNCf033VCvEtH9+C84CwO7ZTJ0X13vZLnktE9yU5L5vEPV/DYjOV43C76dMjkLDsR2ZNxh3Xm3/M28PRnqygqr6V9RjJDu+Xw+v8exckDrVNVt4zrz/qSSn723FxcwISh+dw6bgC3vLEoZvvjpuM7MntdJQ99soX0ZDe+sXlcPab9buNvPqET+dnJ/HV2EXd/sJm0ZBeH5qVw8fB2MWuTiMjBSIlMnGSkeCh8+Iy4br+1ctKSufeswdx71uBmY08d1JnSKRN3W37WsPw9Ji7BS49ocvkVY3tzxdjeTZZFtvP5iMu46/lO7Nvw+vh+HRu1r3deBiWPncX89dt3WV50V+PP2z7dwwsX9thtG6b/onEbLz8il8uPyN1j20VEpHWUyMSJy+XSTRtFRET2kuaREREREcdSl4BIKzV1qklEROJDPTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx9Jg3zgxTZOKiDs2728ZKZ6G+wiJiIg4lRKZOKmoqSP/1nfjtv3Ch8/QPDYiIuJ4OrUkIiIijqV/yRPAucdtI8lj7vPthOpcvPFFh1avVx2q4/EZK/jXvPWsK64kNclDz/bpnDGkC3dOGATAi7MLuOaf39EpK4Vl95xGsmdnjrylrJpB98ygpi7MO9cezfH9dt6Xaf667Tz6wXK+WLmNHVW1dM1J44whXZg8bgCds1MBGHLfDNYWV+6xjQvvPIV/zC7gof/+sNvy3lF37RYREedTIpMAkjwmSa2/9VEbtC1ZmvTaAj5cuoU/nDeE4T1yqagJsWhjKd+sKW4Um5bs4e0FhZw7olvDsn/MLiA/J7VRMvLBks1c8rc5nDUsn39dPYb8nFSWFpZx3ztLOeHxz5jxm+Po0T6dT244nnDYavu6kipOmvI5L145irGH7LxJY8csK+npnpvGp5OOb9Su+nIRETmwKJGRZk3/vpBbTx/AOcN3JieHdc3hwpHdG8X+/KhePP/12l0SmRe+XsvPx/bigXeXNSyrrKnj1//8jhP7d9zlBo8922dwVJ/2jHjwI26YtoBXfzWGThFJSFUoDEBeRjJdctIabd/jcjW5XEREDkwaIyPN6pKTyodLN7O1rLrZ2MuP6slnK7aypqgCgM9XbGVbeQ3nDO+6S9yHy7awubSaG0/t16iOnLRkvMf34b+LN1FSWRubDyEiIgck9chIs566eDi/+vs8Dr3rfQZ0zmJU7/acMrAT543ohse96yXc3XPTOWVgZ174ei13TRjE81+t5adH9iA9eddzZ8s3lwEwOD+7yW0Oys8mbMLKLeUc2Su3xW0tKKkk/5Z3dlnWOTuV7+84pcV1iIiIcyiRkWYddUge3/3uJ8wrKGHW6mJm/ViE96XveOrTVbzrP4a0qCTliqN7Mem1BVxzfB/+s2Ajn1zfeMzKvtI1J413rj1ml2VJHs2XIyJyoNKpJWkRj9vFqN7tufbEvky9YhRvXjOWuWtLmDZvQ6PY8Yd1wQVc9fdvGdatHUO65TSK6dc5E4DFG0ub3N7SwlLcLujbsXVXGiW5XRzaKXOXh65WEhE5cCmRSQChOhehOvbDI3Y9E/WnhLY0MW7G43Zx2ZiefLJ8K1cc3avJ9U8Z2ImOWSk8/uGKRmU7qmoJfv4jpx/WhfYZKTFrs4iIHHh0aikBtGVul/3p9Cdncv6IbozslUuX7FTWlVTy6AfLSfG4GT+kS5Pr3Hr6AK45oc9uE5GMlCSevngEP3tuLldO/YZrT+xLfk4aSzeVct/bS0lJcvP4BcNa3dY602TTjqpGy9tnpJCSpLxdRORAo0RGmjVucGde/24Dj7z/A9srQ3TMSuHIXrm8d90xDOrS9GDdZI+72blbxg/pwozfHsujHyzngr/OorQqRL49Id4tERPitcb6kir63f1Bo+XRE/GJiMiBQYlMnGSkeCh8+Iy4br+lbjy1Pzee2n+PMZeN6cllY3rutrx3XgalUyY2Wn5Ez1xeump0i9uyu3oAbh8/kNvHD2xxXSIi4nxKZOLE5XLppo0iIiJ7SYMGRERExLGUyIiIiIhjKZERERERx1Iis4+5XC5cLmv+llAoFOfWyIEkFDaBMC7CuEx9t0Tk4KTRpvuYy+UiJSWF6upq1q9fT/fu3UlK0m5PVOGwiWmGG16ToHc3CIVNCouqCNdW4AmVJGozRUT2Of1F3Q+6devG2rVrqaqqYuXKlfFujuyBaQK1dQCs2h5K4AQhTLi2AioLabf9k3g3RkQkbpTI7AdpaWn06tWLDRs2UFNTg2ma8W6S7IZpmizbZN3/6bCsbbhJzJ+VizCeUAnttn9Ccqgo3s0REYkbJTL7SVpaGn379sU0TSUyCay8OsQJz/0XgLlHv0i6JzHHnrjMRO4tEhHZf+KayBiB4M3A+cAgrNEIC4EH/D7ve1FxA4AngeOBSmAacIPf5y2LiMkGHrfrSwM+B67z+7zLo+qaDFwL5ANLgFub2N4E4PfAYGAjYPh93j/G4jNHDv6VxON2u6mxzizhMkO4NYhWRCShxfuqpZ8AfwNOBo4CZgFvGYHgsfUBRiCYBXwIhIFjgQuBccDzUXX9HTjNLj/OXjbDCAQzI+q6HrgPuAsYAbwP/McIBI+IiBkFvAn81465G3jACAT9e/9xRUREJJbi2iPj93mjbzZ0oxEInoHVqzLTXnYp0Bm41O/zFgMYgeC1WAlPP7/Pu8LusTkHONPv835kx1wCFAKXAM8YgaALuBl4wu/zvmDXPdkIBE8GbgQus5fdAMzz+7y32O+XGIHgUOAWIxB8yu/zNnteKBwOEw6HW7k3JBHo5xZ/Tjt+nNTWPdF+l9bYF98Xt7ttfSsJNUbGCAQ9QDZQHrH4WGBWfRJjex+rh+Y4YIUdEwYabnvs93mLjUBwth3zDHAI0A3Y5TSS/f7yqO290ETMTUBvYHVzn+PlV6ftcvpo9KiRza0iCaKyVr8c423J0mWkJ8e7s7jlDpTvjPa7tMa++L4MGzqkTesl2rf2DqxE5i8Ry7pi9aw08Pu8tUCRXVYfs81eHqkwKqZ+2e5imtxexPuuiIiISMJImB4ZIxD0AZOBs/0+77p4t2dvXHzRBaSnp8e7GdIG5dUhWtDpJvvQ4EEDHXVn+APlO6P9Lq2RSN+XhGiFEQjeBNwDTKwf4xJhI9AnKj4ZyLPL6mM6GIFgclSvTBdgVUQMWFcrLY6K2RjxfqMdQ1QMUXG75Xa723yuT+JLP7f4c9rx46S27on2u7RGIn1f4t4KIxCsv4rojCaSGLAG/R5lBIK5EctOw2r7FxExbuDUiHpzsa6Eqo9ZDWwATo+qf3xETH1dTcWsA9a04COJiIjIfhLveWT+BHixrixabgSC9T0hNX6ft3660peAO4GXjEDwdiAXMIBpfp93BYDf5/3BCATfBJ4yAsGrgBKseWAKgX/aMaYRCD4KPGQEgouA2cAVWJdY/zqiWVOAL41A8CGsQb9jgOuAm1tyxZKIiIjsP/Hukfkt1uR1/8Y6bVP/eL0+wJ707lSspOtLu+xDrCQk0uX28tftOA9wmt/nLY+o609Yp7AeAOYDZwDn+H3ebyNi5gDnAhPsmPuBu/w+rxGLDywiIiKxE+95ZFo0xa3f512GNQnenmJKgavtx57iHgEeaSbmbeDtlrRNRERE4ifePTIiIiIibaZERkRERBxLiYyIiIg4lhIZERERcSwlMiIiIuJYSmRERETEsZTIiIiIiGMpkRERERHHUiIjIiIijqVERkRERBxLiYyIiIg4lhIZERERcSwlMiIiIuJYSmRERETEsZTIiIiIiGMpkRERERHHUiIjIiIijqVERkRERBxLiYyIiIg4lhIZERERcSwlMiIiIuJYSbGqyAgETwMuBjr7fd6JRiA4Csjx+7wfxWobIiIiIpFi0iNjBILXAU8Dy4ET7MWVwAOxqF9ERESkKbE6tXQ9cKrf530YCNvLlgIDY1S/iIiISCOxSmSygQL7tWk/JwM1MapfREREpJFYJTKfAbdGLfsN8HGM6hcRERFpJFaDfa8DphuB4NVAthEILgNKgbNiVL+IiIhIIzHpkfH7vBuB0cBFwKXAL4Axfp+3MBb1i4iIiDQlJj0yRiA4Atjm93lnA7PtZT2NQDDP7/POj8U2RERERKLFaozMi1iDeyOlAH+PUf0iIiIijcQqkenl93lXRS7w+7wrgUNiVL+IiIhII7FKZNYZgeDIyAX2+w0xql9ERESkkVhdtTQFeNMIBP8ArAQOBW4CHoxR/SIiIiKNxOqqpb8CNwBnAo/azzf6fd6/xKJ+ERERkabE7KaRfp/3X8C/YlWfiIiISHNieffrccAIICtyud/nvStW2xARERGJFKt5ZAysyfA+Bioiisym1xARERHZe7HqkbkUGO73eQuajRQRERGJkVhdfr0VKIlRXSIiIiItEqsemceAfxiB4EPApsiC6InyRERERGIlVonM0/Zz9N2uTcATo22IiIiI7CImiYzf543VKSoRERGRFlMCIiIiIo4Vq8uvkwAfcCLQEXDVl/l93hNisQ0RERGRaLHqkZkCeIHPgCOBaUBn4KMY1S8iIiLSSKwSmfOBM/w+7/8BIfv5XODkGNUvIiIi0kisEpkMoH4yvEojEMzw+7xLgSNiVL+IiIhII7G6/HoJMBqYDcwF7jECwR3A+hjVLyIiItJIrBKZ3wJ19usbsOaVyQb+t7kVjUDwBOBGrBtO9gLu9fu890SUn4R1D6doV/p93ucj4roC/weMtxe9B1zn93k3RcQkAw8ClwO5wLfA9X6fd05Um64AbgMOAVYDD/p93qnNfRYRERHZv2I1j8yciNfLgVNbsXoWsBh4CfjTHuLGsPP0FcD2+hdGIOgG3rLfjrOfA8B0IxAc6/d5w/ayR4GfA1cCK4GbgRlGIHiY3+ddb9d1LvAsMBl4GzgTeM4IBIv9Pu/0VnwuERER2cfanMgYgeBPWhLn93n3eOWS3+d9B3jHrvORPYRu8fu8hbspOxUYCQz1+7yL7LouBxYCP8FKVnKAXwM3+H3eN+2Yq4DTgWuAO+y6JgPT/D7vY/b7pUYgOBa4BWhRIhMOhwmHw80HSsLRzy3+nHb8OKmte6L9Lq2xL74vbnfbhu3uTY/Msy2IMYG+e7GNSJ8YgWAGsAIIAlP9Pq9plx0LFNQnMQB+n3eREQiuA44DZmBdFp6KdcqpPqbOCAQ/sGMwAsEUrLE+z0Rt+z3gaSMQTPb7vLXNNfTlV6fhcjVMpcPoUSNb+1klTipr9csx3pYsXUZ6snPm6jxQvjPa79Ia++L7MmzokDat1+ZExu/z9mnruq20EbgWaxBxGOtUz1+BfsCddkxXoKnemkK7jIjn6LhCrNNWYE3ml7SbmGS7fGNbPoSIiIjEXqwG++4zfp93GbAsYtFcu+dkkhEI3teSHpL97eKLLiA9PT3ezZA2KK8OYY3vlngZPGggmakJ/6upwYHyndF+l9ZIpO9LrG5RkAPcQ9O3KOgVi21E+RK4HegEbMDqJRnfRFwXdvag1D/nA6t2E7MVCNkx0fXU2uXNcrvdbT7XJ/Gln1v8Oe34cVJb90T7XVojkb4vsWpFAGuw7X1AHnAdsBbr1gX7wkigkp2JxUygpxEIDq4PMALBw4CewBf2om+AaqzBvfUxbqyBwl8A+H3eGmBOZIxtPDA7EXt/REREDmax6hcaBwz2+7zbjECwzu/zvmkEgnOxrvLZYzJjBIJZWONdAFKAfCMQHAHU+H3exUYgOAkrKVqENXj4DKzeGMNOPMAazPstMNUIBP32sqewkpKPAPw+7w4jEPwz8IARCK7Huvz6JiATa96ben8AptnbfQeYAFyAdcsFERERSSCx6pFxs3NelzIjEGyHdbqm3+5XaTAKmGc/umLdfHIe9iXZWMnW77ESldlYk9n9ButyaADseWLOwjphOsN+/AicHTGHDFjzxjwL/AWrh2YgcFr9HDJ2XW8Av8S6JHuB/Xyl5pARERFJPLHqkZmPNT7mQ+BzrFNNZcAPza3o93k/IWJMTRPlj2JNZNdcPRuBC5uJqcWaJ2ZyM3HPA883t00RERGJr1j1yFzNzuHjvwWqsG4B8PMY1S8iIiLSSKxuUbAq4vVmrFMzIiIiIvtUzC4Ct6f7vwTohnVJ9MvA3yJm3xURERGJqVjNI/MH4Bysmz6uAXpjXRE0kGbGo4iIiIi0Vax6ZK4ARvp93nX1C4xA8C2sK42UyIiIiMg+EavBvqX2I3rZjhjVLyIiItJIm3tkjEAw8q7WfwJeNwLBh4F1WDPq3sy+m9lXREREZK9OLa3Ammk3cg6Yk6NifgIYe7ENERERkd1qcyLj93kT425RIiIictBSMiIiIiKOpURGREREHEuJjIiIiDiWEhkRERFxrH2SyBiB4MlGIHjivqhbREREpF5MEhkjEPzUCASPtV/fgnWfpZeMQPD2WNQvIiIi0pRY9cgMBb62X1+NNZ/MWODXMapfREREpJFY3WvJDZhGIHgo4PL7vIsBjECwfYzqFxEREWkkVonMF1gz+HYF/g1gJzVbY1S/iIiISCOxOrV0BVACfA/cbS8bBPxfjOoXERERaSQmPTJ+n3cbcHvUsrdjUbeIiIjI7sQkkTECwRuAj/w+73dGIDgWeBWoAy71+7xfxWIbIiIiItFidWppEvCj/foh4HHgAeBPMapfREREpJFYJTLt/D7vdiMQzAaGA0/6fd5ngYExql9ERESkkVhdtVRgBILHAEOAz/w+b50RCOZgnV4SERER2SdilcjcDLwG1AAX2MvOAmbHqH4RERGRRmJ11dI7QLeoxf+yHyIiIiL7RKx6ZACwx8h0BFwRi1fFchsiIiIi9WJ1+fVhwD+wBvqaWImMaRd7YrENERERkWixumopAHwM5AE7gPZAEPhFjOoXERERaSRWicxw4Ba/z1uCddPI7VgDgO+PUf0iIiIijcQqkakCku3XW41AsJddd4cY1S8iIiLSSKwSmc+Bi+zXrwHvAp8CH8WofhEREZFGYnX59UURb28HFgFZwNRY1C8iIiLSlJhefg3g93nDwN9jXa+IiIhItDYnMkYg+Hd2XmK9W36f9+dt3YaIiIjInuxNj8yKmLVCREREpA3anMj4fd57Y9kQERERkdbaq6uWjEDwGCMQfGQ3ZQ8bgeDYvalfREREZE/29vLr3wGf7absU7tcREREZJ/Y20RmBPDebso+AI7cy/pFREREdmtvE5kcIGU3ZclA9l7WLyIiIrJbe5vILAXG7aZsnF0uIiIisk/s7YR4U4CgEQh6gDf8Pm/YCATdwLnAU8ANe1m/iIiIyG7tVY+M3+d9CfgD8AJQZQSCG7BuIPkC8Kjf5/3n3jdRREREpGl7fdNIv8/7ONAdmAjcZD93t5eLiIiI7DOxumnkDuC/sahLREREpKX2ukdGREREJF6UyIiIiIhjKZERERERx4rJGJm9YQSCJwA3Ys0S3Au41+/z3hMVMwB4EjgeqASmATf4fd6yiJhs4HHgfCAN+By4zu/zLo+qazJwLZAPLAFu9fu870XFTAB+DwwGNgKG3+f9Y2w+sYiIiMRKIvTIZAGLgclAYXShEQhmAR8CYeBY4EKsyfaejwr9O3CaXX6cvWyGEQhmRtR1PXAfcBdW4vQ+8B8jEDwiImYU8CbW4OURwN3AA0Yg6N+bDykiIiKxF/ceGb/P+w7wDsBu7qR9KdAZuNTv8xbbcdcCbxmBYD+/z7vC7rE5BzjT7/N+ZMdcgpUYXQI8YwSCLuBm4Am/z/uCXfdkIxA8GatH6DJ72Q3APL/Pe4v9fokRCA4FbjECwaf8Pq/Z3GcKh8OEw+FW7glJBPq5xZ/Tjh8ntXVPtN+lNfbF98XtblvfStwTmRY4FphVn8TY3sfqoTkOWGHHhLFuVAmA3+ctNgLB2XbMM8AhQDca3+TyPeDyqO290ETMTUBvYHVzDX751Wm4XK6G96NHjWxuFUkQlbX65RhvS5YuIz05ETqLW+ZA+c5ov0tr7Ivvy7ChQ9q0nhO+tV2JOuXk93lrgSK7rD5mm708UmFUTP2y3cU0ub2I910RERGRhOGEHhnHufiiC0hPT493M6QNyqtDtKDTTfahwYMGkpnqnF9NB8p3RvtdWiORvi+J0Yo92wj0iVxgBILJQJ5dVh/TwQgEk6N6ZboAqyJiwLpaaXFUzMaI9xvtGKJiiIrbLbfb3eZzfRJf+rnFn9OOHye1dU+036U1Eun7khit2LOZwFFGIJgbsew0rLZ/ERHjBk6tD7Djj4qIWQ1sAE6Pqn98REx9XU3FrAPWtO0jiIiIyL4Q9x4Z+/LqfvbbFCDfCARHADV+n3cx8BJwJ/CSEQjeDuQCBjDN7/OuAPD7vD8YgeCbwFNGIHgVUII1D0wh8E87xjQCwUeBh4xAcBEwG7gC6xLrX0c0aQrwpREIPoQ16HcMcB1wc0uuWBIREZH9JxF6ZEYB8+xHV8Brv34HwJ707lSspOtL4HWseWWuiKrncnv563acBzjN7/OW1wf4fd4/AfcADwDzgTOAc/w+77cRMXOAc4EJdsz9wF1+n9eI1QcWERGR2Ih7j4zf5/0EcDUTswxrErw9xZQCV9uPPcU9AjQ1X01kzNvA23uKERERkfhLhB4ZERERkTZRIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYyXFuwEtYQSC9wB3N1HUx+/zrrZjjgKmACOBYmAq8Du/zxuKqKcr8H/AeHvRe8B1fp93U0RMMvAgcDmQC3wLXO/3eefE9EOJiIjIXnNSj8w6oGvUowDACAR7Ah8Ay4FRwK+Bq4GH61c2AkE38BZwKDDOfvQDpttl9R4FfmXXMRpYAcwwAsHu+/CziYiISBs4okfGVuf3eQt3U3YNUA5c5fd564CFRiB4J/BHIxC81+/zlgKnYvXWDPX7vIsAjEDwcmAh8BOsZCUHK4G5we/zvmnHXAWcbm/jjpY0NBwOEw6H2/o5JY70c4s/px0/Tmrrnmi/S2vsi++L2922vhUnJTL5RiBYALiABcD9fp/3S7vsWOADO4mp9x5gAEcCn9gxBfVJDIDf511kBILrgOOAGXZsqr1ufUydEQh+YMe0yMuvTsPlcjW8Hz1qZCs+psRTZa1+OcbbkqXLSE92TmfxgfKd0X6X1tgX35dhQ4e0aT2nfGtnA1cCZwGXYI2B+dwIBE+zy7sC0b01hRFlu4upj+saFdtUXV0RERGRhOKIHhm/z/tO5HsjEPwC6AXcjDU2JqFcfNEFpKenx7sZ0gbl1SFgdbybcVAbPGggmamO+NUEHDjfGe13aY1E+r4kRitaye/zmkYg+BVwjr1oI5AfFdYloqz+eTyNdYmKwa5r1W5imuV2u9t8rk/iSz+3+HPa8eOktu6J9ru0RiJ9XxKjFW0zEvuqJWAmcGrU1UfjgSrgm4iYnkYgOLg+wAgEDwN6Al/Yi74BqrEG99bHuLEGCtfHiIiISIJwRI+MEQg+jnXp9GogB/ACJ7OzR+ZpwA88awSCf8S6xPp+4Cn7iiWwBvN+C0w1AkG/vewpYA7wEYDf591hBIJ/Bh4wAsH1wErgJiDT3oaIiIgkEKf0yHTFmuBuCfA+0B841e/zTgfw+7wFWPPCDMTqVfkL8Axwa30Ffp83jDVYeDVWUjMD+BE42y6rdzPwrF3HN3adp/l93vX77uOJiIhIWziiR8bv817SgpivgWOaidkIXNhMTC0w2X6IiIhIAnNKj4yIiIhII0pkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjKZERERERx1IiIyIiIo6lREZEREQcS4mMiIiIOJYSGREREXEsJTIiIiLiWEpkRERExLGUyIiIiIhjJcW7AdJ2q7aWc9O0BcxZU0J6iodrju/DpFP6UVBcweiHP9kltioUZtzgzrz6qzFN1rWjqpbr/7WA9xZtIi3Zjff4PtwybkCLyw+mtoscTHSsSqJTIuNQdWGTnz4zm7OG5fPKr8bw47YKznn6K7rnpnHRkT0ofGRCQ2xNKEz/e97nf47ottv6bpq2kOKKGhbffSpbSqs5++mv6Nk+nUtH92xR+cHSdpGDiY5VcQKdWnKo5ZvLWL6lnNtOH0iyx82Azln8/KhePPfV2kaxby0oxDTh7MO7NllXRU2IafM2cOcZg8hNT6Z/5yy8x/dh6qy1LSo/mNoucjDRsSpOoETGocKmCYCJGbEMFm3c0Sh26qy1XDSyO2nJnibrWr65nJq6MId3z2lYdnj3dizaUNqi8oOp7SIHEx2r4gRKZByqf+cseuel88C7y6gO1bFkYyl/n72WHVWhXeLWFlXw8Q9b+MXYXrutq6w6RGaKhyTPzq9Du/RkSqtDLSo/mNoucjDRsSpOoETGoZI9bl6+agzfr9vOgHs+4JcvfstlY3qSl5G8S9yLswsY3r0dw7q3221dWalJVNTWEaoLNyzbUVlLdmpSi8oPpraLHEx0rIoT6BvSBCMQnAD8HhgMbAQMv8/7x/3ZBtM0qaip22NMr7x0XrpqdMP7+99dytg+eZTb/8GEwyZTZ63lNycd2rCsKd3apZHsdjF7dTHDe1i/iOauLWZQlyzKq0PNljfZ9lqr7RnJHlwul2PaXl6j//7k4LS7776OVUl0LtM0m486iBiB4CjgK+Bx4HlgDBAEbvL7vEYT8Z2BTZHLLr3kItLT0vaqHRXVIYY88OEeY8Im1KcIdUB12E26O4zbXhgyrWUZ7jBN5BK7qA67ME0Xqe4wJlAVdpPsMkl2my0qby0ntP3TMS+T7t5zMplIimpTGf/N/wAw8egiPE0PVUhIdXUw/as8ABbdcQoZDvovvCXHqpPpWI0tHae79+xzUzvbL7f5fd7wHoMjKJGJYgSCLwH9/D7vmIhljwIXA738Pq8ZFT8IWLJ/W2l5f2s2X2/PJGRC19QQZ3TcwSHpNQ3lL21sT5LL5KL8kkbrPrc+j0PSazg5rwyAqjoX/97cjqXlaSS7TY5uV84pHcoa4psrP5jaLnIw0bEqcdDZ7/NuaWmwEpkoRiC4BnjB7/PeFbHsFGAG0Mfv866Oio9bIiMiInIAalUio8G+jXUFCqOWFUaUiYiISIJQIiMiIiKO5ZwRdfvPRiA/almXiLJoy7GubopUBOicnXN9bj8fH9dWiEhzdKwemLa1JliJTGMzgdOBuyKWjQfWAWuig/0+bx2wdP80TfaTQfFugIi0iI5VUSLThCnAl0Yg+BDwAtbl19cBN0dfsSQiIiLxpauWmmAEgmdiTYg3CGugr+H3eR+Nb6tEREQkmhIZERERcSxdtSQiIiKOpURGREREHEuJjIiIiDiWEhk5qBiBYC8jECwzAkGP/f4TIxD8lf36CiMQ/KKN9bZ5XRFpzAgE3zUCwV/sp23p+HUwXX4tByQjEFwN/Mrv886w318MPA2c6/d5s+LZNpGDWfSxuTt+n/eM/dMicTr1yMgBz/6v7ingTL/P+2m82yMi+44RCLqMQFB/2w4i6pGRA5oRCHqBB4HT/T7vXCMQPAT4EUj2+7yhZtYdBDwJHAlsAe70+7yv2mUdgOeAk7Bmdv7vvvoMIgciIxC8AvgV8DXwS6AE8Pl93nft8k+AF/0+7zNGIHgP0M/v815mlx1CxHFsx87EOh5HAsOMQDAJHb8HBWWtciC7BrgPOMXv885tzYpGIJgJfAC8BHQGLgYCRiB4mB3yFFCFdUf0q+yHiLTOUcAyoCPwB+BZIxB0tbGuy4H/BbKxEhcdvwcJ9cjIgew04GNgQRvWPQtY7fd5n7PfzzMCwWnAhUYg+ABwATDM7/OWAwuNQPAF4IRYNFrkILLG7/P+FcA+hgJYN+ktbENdz/t93kV2XePR8XvQUCIjB7JrgDuAZ4xA8JetvFdWb+AoIxAsiViWBPwd6GS/Logoa3RDURFpVkPC4vd5K4xAEKCtg/Ejj0cdvwcRJTJyINsEnAJ8ivWf3jWtWLcA+NTv854WXWBfuh0CerLzzue99q6pIrIH5UBGxPv8JmIi/1HR8XsQUSIjBzS/z7vBCARPAT41AsEpwP+1cNW3gIeNQPBy4GV72QigzO/zLjECwdeBe4xA8CrgEOAXwOpYtl1EGnwH3GIEgr2A7cBtzcTr+D2IaLCvHPD8Pu9a4CfA/wAPtXCdUmAc1iDBDVhd4I8AqfUhWF3ghcDzWFdAiMg+4Pd5PwBeAb4HvsFKVPYUr+P3IKK7X4uISMIxAsHPgGf8Pu/UeLdFEpt6ZEREJKEYgWAG0BdrrhiRPVIiIyIiCcMIBDtjnfL5FND9j6RZOrUkIiIijqUeGREREXEsXX4tIvudEQgOxLoK5VDgd36f94k4N0lEHEqJjIjEw2TgY7/POyLeDRERZ9OpJRGJh97AoqYK7JlXRURaRIN9RWS/MgLBj4ATgVqsqeL/gzVba297+TnAYuBJrBv5lQFT6k8/GYFgOvC0HbcRazKz3/p93h52uQn09/u8K+z3zwPr/D7vHfb7s4AHsGZ0XQz82u/zfm+XrQYM4Od2e94DfuH3eavs8nOAe7EuDd4CXIt1t+Vb/T7vkRGf8QbgRL/Pe07MdpyINEk9MiKyX/l93p8AnwN+v8+bBdQAlwIPYiUFXwLTgflAd6z7ZV1vBIKn21XcjTW25lDgdKzp5VvECASPAP4GeIEOQBD4jxEIpkaEXQSMB/oAhwNX2OuOAaYCNwO5WEnWaqxErI8RCA6OqONyO1ZE9jGNkRGRRPCm3+edCWAEgsOATn6f9z67bJURCP4Va7r5/2IlGj6/z1sEFBmB4BPAXS3czv8CQb/PO8t+/4IRCN4OjMWatwTgCb/Pu8Fuy3Sse/QA/BL4mz1dPsD6+kqNQPAV4DLgd0YgOASrt2eP0+iLSGwokRGRRFAQ8bo30M0IBEsilnmwenEAukXFr2nFdnoDvzACwesilqXYddYrjHhdEVHWE3hnN/W+APzTCATvwOqNedXv81a3ol0i0kZKZEQkEUQO1isAfvT7vP13E7sRK6moHyzcK6q8AsiIeJ8PrIuo+0G/z/tgG9pYgHU6qxG/z/u1EQjWAMdjnSa7tA31i0gbKJERkUQzGyg1AsFbgCewxtAMBtL9Pu8c4FXgNiMQnAVkAtdFrf8dcKkRCC4CTsMaQDzXLvsr8G8jEJxhbycDOAn4zL5j8p48C7xvBIJvAR8DXYFsv8+71C6fijVQuNbv82pqfZH9RIN9RSSh+H3eOuAsrLEpPwJbgWeAdnbIvVink34E3gf+HlXFb4GJQAnwM+CNiLrnAldjJRzFwArswbwtaNds4EpgCtZVVp9inaqq93dgKPBiS+oTkdjQ5dci4mhGIHgS8GL95ddxbEc6sBkY6fd5l8ezLSIHE/XIiIjExjXAHCUxIvuXxsiIiOwleyI9F3BufFsicvDRqSURERFxLJ1aEhEREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxLCUyIiIi4lhKZERERMSxlMiIiIiIYymREREREcdSIiMiIiKOpURGREREHEuJjIiIiDiWEhkRERFxrP8HirPscc44qpEAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 600x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.pyplot import figure, show\n",
    "from libs.ds_charts import multiple_bar_chart\n",
    "\n",
    "figure()\n",
    "multiple_bar_chart([positive_class, negative_class], values, title='Target', xlabel='frequency', ylabel='Class balance')\n",
    "show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## For zscale\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 97\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 0.0 : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "original = read_csv(filename_zscore, sep=',', decimal='.')\n",
    "# drop this because later, knn does not handle not float values.\n",
    "#original.drop(['CRASH_DATE','CRASH_TIME'],axis=1,inplace=True)\n",
    "class_var = 'PERSON_INJURY'\n",
    "target_count = original[class_var].value_counts()\n",
    "positive_class = target_count.idxmin()\n",
    "negative_class = target_count.idxmax()\n",
    "#ind_positive_class = target_count.index.get_loc(positive_class)\n",
    "print('Minority class=', positive_class, ':', target_count[positive_class])\n",
    "print('Majority class=', negative_class, ':', target_count[negative_class])\n",
    "print('Proportion:', round(target_count[positive_class] / target_count[negative_class], 2), ': 1')\n",
    "values = {'Original': [target_count[positive_class], target_count[negative_class]]}\n",
    "\n",
    "figure()\n",
    "bar_chart(target_count.index, target_count.values, title='Class balance')\n",
    "savefig(f'images/{file_zscore}_balance.png')\n",
    "show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_positives = original[original[class_var] == positive_class]\n",
    "df_negatives = original[original[class_var] == negative_class]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 97\n",
      "Majority class= Injured : 97\n",
      "Proportion: 1.0 : 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import concat, DataFrame\n",
    "\n",
    "df_neg_sample = DataFrame(df_negatives.sample(len(df_positives)))\n",
    "df_under = concat([df_positives, df_neg_sample], axis=0)\n",
    "df_under.to_csv(f'data/dataset_1/{file_zscore}_under.csv', index=False)\n",
    "values['UnderSample'] = [len(df_positives), len(df_neg_sample)]\n",
    "print('Minority class=', positive_class, ':', len(df_positives))\n",
    "print('Majority class=', negative_class, ':', len(df_neg_sample))\n",
    "print('Proportion:', round(len(df_positives) / len(df_neg_sample), 2), ': 1')\n",
    "# here we split by the minory getting 97 of each"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 27220\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 1.0 : 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import concat, DataFrame\n",
    "\n",
    "df_pos_sample = DataFrame(df_positives.sample(len(df_negatives), replace=True))\n",
    "df_over = concat([df_pos_sample, df_negatives], axis=0)\n",
    "df_over.to_csv(f'data/dataset_1/{file_zscore}_over.csv', index=False)\n",
    "values['OverSample'] = [len(df_pos_sample), len(df_negatives)]\n",
    "print('Minority class=', positive_class, ':', len(df_pos_sample))\n",
    "print('Majority class=', negative_class, ':', len(df_negatives))\n",
    "print('Proportion:', round(len(df_pos_sample) / len(df_negatives), 2), ': 1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Minority class= Killed : 27220\n",
      "Majority class= Injured : 27220\n",
      "Proportion: 1\n"
     ]
    }
   ],
   "source": [
    "from pandas import Series\n",
    "from imblearn.over_sampling import SMOTE\n",
    "RANDOM_STATE = 42\n",
    "\n",
    "smote = SMOTE(sampling_strategy='minority', random_state=RANDOM_STATE)\n",
    "y = original.pop(class_var).values\n",
    "X = original.values\n",
    "smote_X, smote_y = smote.fit_resample(X, y)\n",
    "df_smote = concat([DataFrame(smote_X), DataFrame(smote_y)], axis=1)\n",
    "df_smote.columns = list(original.columns) + [class_var]\n",
    "df_smote.to_csv(f'data/dataset_1/{file_zscore}_smote.csv', index=False)\n",
    "\n",
    "smote_target_count = Series(smote_y).value_counts()\n",
    "values['SMOTE'] = [smote_target_count[positive_class], smote_target_count[negative_class]]\n",
    "print('Minority class=', positive_class, ':', smote_target_count[positive_class])\n",
    "print('Majority class=', negative_class, ':', smote_target_count[negative_class])\n",
    "print('Proportion:', round(smote_target_count[positive_class] / smote_target_count[negative_class]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.pyplot import figure, show\n",
    "from libs.ds_charts import multiple_bar_chart\n",
    "\n",
    "figure()\n",
    "multiple_bar_chart([positive_class, negative_class], values, title='Target', xlabel='frequency', ylabel='Class balance')\n",
    "show()"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
  },
  "kernelspec": {
   "display_name": "Python 3.10.0 64-bit",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.0"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}