car_dataset_price_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_pred(x, w, b):\n",
    "    return np.matmul(x, w) + b\n",
    "\n",
    "def compute_cost(x, y, w, b):\n",
    "    sqr_diff = (make_pred(x, w, b) - y)**2\n",
    "    return sqr_diff.sum() / (2*x.shape[0])\n",
    "\n",
    "def run_grad_decent(x, y, w, b, alpha):\n",
    "    cost_sum = ((make_pred(x, w, b)-y).sum() / x.shape[1])\n",
    "    new_b = b - alpha * cost_sum\n",
    "\n",
    "    j = x.shape[1]\n",
    "    i = x.shape[0]\n",
    "\n",
    "    preds = x*(np.matmul(x, w) + b - y)\n",
    "    new_w = np.sum(preds, axis=0)\n",
    "\n",
    "    new_w /= i\n",
    "    new_w = w - (alpha * new_w.reshape([j, 1]))\n",
    "\n",
    "    return new_w, new_b\n",
    "\n",
    "def run_linear_regression(x, y, alpha = .004, iter = int(1e3)*2):    \n",
    "    update_alpha = True\n",
    "    j = x.shape[1]\n",
    "    w = np.zeros([j, 1])\n",
    "    b = 0\n",
    "    cost_history = []\n",
    "    print(b)\n",
    "    cost_history.append(compute_cost(x, y, w, b))\n",
    "    print(cost_history[-1])\n",
    "    for _ in range(iter):\n",
    "        new_w, new_b = run_grad_decent(x, y, w, b, alpha)\n",
    "        cost_history.append(compute_cost(x, y, new_w, new_b))\n",
    "        float('nan')\n",
    "        if math.isinf(cost_history[-1]) or math.isnan(cost_history[-1]) or cost_history[-1] >= cost_history[-2]:\n",
    "            alpha -= .001\n",
    "            update_alpha = False\n",
    "            new_w, new_b = w, b\n",
    "            cost_history.pop()\n",
    "        elif update_alpha:\n",
    "            alpha += .001\n",
    "        w, b = new_w, new_b\n",
    "    print(compute_cost(x, y, w, b))\n",
    "    return w, b, cost_history "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1.4051423544996315\n",
      "0.9044157801171571\n",
      "0\n",
      "1.4051423544996315\n",
      "0.6602166166301597\n",
      "0\n",
      "1.4051423544996315\n",
      "0.40304752316970993\n",
      "0\n",
      "1.4051423544996315\n",
      "0.35322532352672364\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGdCAYAAADAAnMpAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA+4UlEQVR4nO3deXhUhb2H8e/MJDOTkA0IJCEJ+yZbwEBCtFbbRvHqtdrayiZLIKBetVp6a6EuXLWV9rZFbaWVfROEut9WiwsuFSUJJCBB9p1AFgIkk4VsM+f+EYxGAmRCkpPl/TzPPH08nDPzy3nmmXk7Z+Yci2EYhgAAAExiNXsAAADQvhEjAADAVMQIAAAwFTECAABMRYwAAABTESMAAMBUxAgAADAVMQIAAEzlY/YA9eHxeHTy5EkFBgbKYrGYPQ4AAKgHwzBUVFSkbt26yWq9+OcfrSJGTp48qejoaLPHAAAADXD8+HFFRUVd9N9bRYwEBgZKqv5jgoKCTJ4GAADUh8vlUnR0dM37+MW0ihj56tBMUFAQMQIAQCtzua9Y8AVWAABgKmIEAACYihgBAACmIkYAAICpiBEAAGAqYgQAAJiKGAEAAKYiRgAAgKmIEQAAYCpiBAAAmIoYAQAApiJGAACAqRoUIwsWLFDPnj3ldDoVHx+vtLS0i65bWVmpp556Sn369JHT6VRMTIw2bNjQ4IEb06f7Tyl55Radq3CbPQoAAO2W1zGyfv16zZo1S3PnzlVGRoZiYmI0ZswY5eXl1bn+Y489poULF+ovf/mLdu3apXvvvVc/+tGPtG3btise/kqUVlTp4XXb9cHuPCWvIkgAADCLxTAMw5sN4uPjNWrUKL3wwguSJI/Ho+joaD344IOaPXv2Bet369ZNjz76qO6///6aZXfeeaf8/Pz00ksv1esxXS6XgoODVVhYqKCgIG/GvaQtR85o6rI0lVS4ldC7s5ZOHSl/u0+j3T8AAO1Zfd+/vfpkpKKiQunp6UpMTPz6DqxWJSYmavPmzXVuU15eLqfTWWuZn5+fNm3adNHHKS8vl8vlqnVrCqN6dtLKaXEKcPho86HTSlq+RaUVVU3yWAAAoG5exUh+fr7cbrfCwsJqLQ8LC1NOTk6d24wZM0bz58/X/v375fF49P777+v1119Xdnb2RR9n3rx5Cg4OrrlFR0d7M6ZXRn4jSFIPn9HUZVtUUk6QAADQXJr81zTPP/+8+vXrp4EDB8put+uBBx5QUlKSrNaLP/ScOXNUWFhYczt+/HiTzhjbo6NWTY9ToMNHaUfOaOryNBUTJAAANAuvYiQ0NFQ2m025ubm1lufm5io8PLzObbp06aI333xTJSUlOnr0qPbs2aOAgAD17t37oo/jcDgUFBRU69bUru7eUauT4xXo9NGWI2c1ZVmaisoqm/xxAQBo77yKEbvdrtjYWG3cuLFmmcfj0caNG5WQkHDJbZ1OpyIjI1VVVaXXXntNt99+e8MmbkLDo0O0JjleQU4fpR89q8nL0uQiSAAAaFJeH6aZNWuWFi9erJUrV2r37t267777VFJSoqSkJEnS5MmTNWfOnJr1U1NT9frrr+vQoUP69NNPdfPNN8vj8eiRRx5pvL+iEQ2LCtHaGaMV7OerbccKNHkpQQIAQFPyOkbGjh2rP/7xj3riiSc0fPhwbd++XRs2bKj5UuuxY8dqfTm1rKxMjz32mAYNGqQf/ehHioyM1KZNmxQSEtJof0RjGxIZrDXJ8Qrx99X24wWatCRVhecIEgAAmoLX5xkxQ1OdZ+Rydp10aeKSFJ0trdSwqGCtnhavYH/fZnt8AABasyY5z0h7M6hbkNbOGK1OHezakVWoiUtTVFBaYfZYAAC0KcTIZVwVEaSXZ4xW5w527Tzh0oTFqTpbQpAAANBYiJF6GBAeqJdnjlZogF27sl2asCRVZwgSAAAaBTFST/3DAvXyjNEKDXBod7ZLExan6HRxudljAQDQ6hEjXugXFqh1M0erS6BDe3KKNGFxqvIJEgAArggx4qW+XQO0buZodQ10aG9ukcYvStGpIoIEAICGIkYaoE+X6iAJC3Jof16xxi9OUV5RmdljAQDQKhEjDdS7S4DWzUxQeJBTB/KKNX5RivJcBAkAAN4iRq5Ar9AOWn/PaHULdurgqRKNW5SiXIIEAACvECNXqEfnDlo3M0GRIX46lF8dJDmFBAkAAPVFjDSC7p39tW7maEWG+OlwfonGLdqs7MJzZo8FAECrQIw0kuhO1UES1dFPR06XauzCFJ0oIEgAALgcYqQRRXfy1/p7EhTdyU/HzpRq3KLNyjpbavZYAAC0aMRII4sM8dP6mQnq0dlfx8+c07hFKTp+hiABAOBiiJEm0C3ET+tmjlbPzv7KOkuQAABwKcRIE4kI9tO6mQnqHdpBJwrOaezCzTp2miABAODbiJEmFB7s1MszR6t3lw46WVimsYs260h+idljAQDQohAjTSwsyKl1M0arT5cOyi4s07hFKTpMkAAAUIMYaQZdg5xaNzNB/boGKMdVpnGLNuvgqWKzxwIAoEUgRppJl0CHXp45Wv3DApTrKtf4RSk6kEeQAABAjDSj0ACHXp4xWgPDA5VXVK5xi1J0IK/I7LEAADAVMdLMOgc4tPZ8kOQXVwfJ/lyCBADQfhEjJujUwa6XZ4zWoIgg5RdXaNyiFO3NIUgAAO0TMWKSjh3sWjsjXoO7Bel0SYXGL07RnhyX2WMBANDsiBEThfjbtSY5XkMjg3WmpELjF6Vo10mCBADQvhAjJgvxt+ul6fGKiQrW2dJKTViSoi9PFpo9FgAAzYYYaQGC/X21anq8YqJDVFBaqQmLU7XzBEECAGgfiJEWItjPV6unx2lE9xAVnqvUhMUpyswiSAAAbR8x0oIEOX21alqcru4eIldZlSYuSdEXxwvMHgsAgCZFjLQwgc7qQzYje3SUq6xKdy9N1bZjZ80eCwCAJkOMtEABDh+tmBanuJ6dVFRWpclL05RBkAAA2ihipIUKcPhoedIoxfXqpKLy6iBJP3rG7LEAAGh0xEgL1sHhoxVJozS6dycVnw+SrUcIEgBA20KMtHD+dh8tnxqna/p0VkmFW5OXpSntMEECAGg7iJFWwM9u09Ipo/SdvqEqrXBr6vI0pRw6bfZYAAA0CmKklfCz27Rkykhd1686SJKWb9HmgwQJAKD1I0ZaEaevTYsnj9R3+3fRuUq3klak6fMD+WaPBQDAFSFGWhmnr02LJsXqhgFdVFbpUdKKLdq0nyABALRexEgr5PS1aeGkWH1/YFeVV3k0feUW/XvfKbPHAgCgQYiRVsrhY9Pf7r5aiVdVB0nyqq36eG+e2WMBAOA1YqQVc/jY9NeJsbpxUJgqqjyauSpdH+0hSAAArQsx0srZfaxaMOFqjRkcpgq3R/esTtfG3blmjwUAQL0RI22A3ceqFyZcrf8YEq4Kt0f3vpSuD3YRJACA1oEYaSN8bVb9efwI3To0QpVuQ/etSdd7X+aYPRYAAJdFjLQhvjarnh83XP85rDpI/mtNhjbsJEgAAC0bMdLG+Nisem7scP0wppuqPIYeWJuhf2Vmmz0WAAAXRYy0QT42q+bfFaM7hp8Pkpe36e0dBAkAoGUiRtooH5tVf7pruH48IlJuj6Gfrdumf3xx0uyxAAC4ADHShtmsFv3hpzH6SWyU3B5DD63bpre2nzB7LAAAaiFG2jib1aL/vXOY7hoZJY8h/Xz9dr25jSABALQcDYqRBQsWqGfPnnI6nYqPj1daWtol13/uuec0YMAA+fn5KTo6Wj//+c9VVlbWoIHhPavVot/9eJjGjYqWx5Bm/X27Xs/IMnssAAAkNSBG1q9fr1mzZmnu3LnKyMhQTEyMxowZo7y8uk9DvnbtWs2ePVtz587V7t27tXTpUq1fv16//vWvr3h41J/VatEzPxqq8XHd5TGkX7zyhV5NJ0gAAObzOkbmz5+vGTNmKCkpSYMGDdKLL74of39/LVu2rM71P//8c1177bWaMGGCevbsqZtuuknjx4+/7KcpaHxWq0W/vWOI7h7dXYYh/fLVL/T3LcfNHgsA0M55FSMVFRVKT09XYmLi13dgtSoxMVGbN2+uc5trrrlG6enpNfFx6NAhvfPOO7rlllsu+jjl5eVyuVy1bmgcVqtFT98+RJMTesgwpEde26F1acfMHgsA0I75eLNyfn6+3G63wsLCai0PCwvTnj176txmwoQJys/P13e+8x0ZhqGqqirde++9lzxMM2/ePD355JPejAYvWCwWPfnDwbJaLFrx+RHNfj1THkOaEN/d7NEAAO1Qk/+a5uOPP9Yzzzyjv/71r8rIyNDrr7+ut99+W08//fRFt5kzZ44KCwtrbsePcyihsVksFs29bZCSru0pSfr1G5l6KeWouUMBANolrz4ZCQ0Nlc1mU25u7SvC5ubmKjw8vM5tHn/8cU2aNEnJycmSpKFDh6qkpEQzZ87Uo48+Kqv1wh5yOBxyOBzejIYGsFgseuI/B8lqsWjppsN67M2d8hiGJif0NHs0AEA74tUnI3a7XbGxsdq4cWPNMo/Ho40bNyohIaHObUpLSy8IDpvNJkkyDMPbedHILBaLHrv1Ks38bm9J0hNvfakVnx02eSoAQHvi1ScjkjRr1ixNmTJFI0eOVFxcnJ577jmVlJQoKSlJkjR58mRFRkZq3rx5kqTbbrtN8+fP14gRIxQfH68DBw7o8ccf12233VYTJTCXxWLRnP8YKItFWvjJIf3PP3bJY0jTvtPL7NEAAO2A1zEyduxYnTp1Sk888YRycnI0fPhwbdiwoeZLrceOHav1Schjjz1W/f++H3tMJ06cUJcuXXTbbbfpt7/9beP9FbhiFotFs28eKJvFor9+fFBP/XOXPIah5Ot6mz0aAKCNsxit4FiJy+VScHCwCgsLFRQUZPY4bZphGPrTe/v0wkcHJEm/vmWgZn63j8lTAQBao/q+f3NtGtRisVj0i5v662c/6CdJeuadPXrxk4MmTwUAaMuIEVzAYrFo1o399XBidZD87l97tOD8JyUAADQ2YgQX9XBif826sb8k6Q/v7tULH+43eSIAQFtEjOCSfvaDfvrvm6qD5I/v7dOfNxIkAIDGRYzgsh74fj89cvMASdL89/fpuQ/2mTwRAKAtIUZQL/91Q1/N/o+BkqTnPtiv+e/v46R1AIBGQYyg3u69vo8eveUqSdKfN+7Xn94jSAAAV44YgVdmfLe3Hru1Okhe+OiA/vDuXoIEAHBFiBF4Lfm63nriPwdJkv768UH9bsMeggQA0GDECBpk2nd66ckfDpZUfT2bef8iSAAADUOMoMGmXNNTT99eHSSL/n1Iv3l7N0ECAPAaMYIrMimhp35zxxBJ0tJNh/XUP3cRJAAArxAjuGJ3j+6hZ340VJK0/LMjevIfBAkAoP6IETSKCfHd9fs7h8pikVZ8fkRPvPUlQQIAqBdiBI1m7Kju+v2dw2SxSKtTjuqxN3fK4yFIAACXRoygUd01Mlp/+EmMLBZpTeoxPUqQAAAugxhBo/tJbJT+9NMYWS3Sy2nH9Os3MgkSAMBFESNoEj++Okrz7xouq0Vat+W4fvXaDoIEAFAnYgRN5o4RkXp2bHWQvJKepV++ukNuggQA8C3ECJrU7cMj9fy4EbJZLXotI0u/fOULggQAUAsxgiZ3W0w3/fl8kLy+7YR+8fftBAkAoIaP2QOgfbh1WISsFunBl7fpze0n5TGk+XfFyMdGDwNAe8c7AZrNfwyN0AsTrpaP1aL/++KkHl6/XVVuj9ljAQBMRoygWd08JFx/nXi1fG0W/XNHth5at12VBAkAtGvECJrdTYPD9beJsfK1WfR2ZrZ+9vI2ggQA2jFiBKZIHBSmhZNiZbdZ9a+dOXpgbYYqqggSAGiPiBGY5vsDw7RwcqzsPla9+2Wu7idIAKBdIkZgqu8N6KrFk0fK7mPV+7ty9V9r0lVe5TZ7LABAMyJGYLrr+3fRkskj5fCx6oPdebrvpQyCBADaEWIELcJ3+3fR0imj5PS16sM9ebpndbrKKgkSAGgPiBG0GN/pF6pl54Pk472nNJMgAYB2gRhBi3JN31AtnxonP1+b/r3vlGas2kqQAEAbR4ygxUno01krkkbJ327Tp/vzlbxyq85VECQA0FYRI2iR4nt31oqkOHWw27TpQL6mr9yi0ooqs8cCADQBYgQtVlyvTlo5rTpIPj94WtNWECQA0BYRI2jRRvbspFXT4xTg8FHKoTOaunyLSsoJEgBoS4gRtHixPaqDJNDho7TDZzR1eZqKCRIAaDOIEbQKV3fvqNXJ8Qp0+mjLkbOauixNRWWVZo8FAGgExAhajeHRIVqTHK8gp4+2Hj2rKQQJALQJxAhalWFRIVqTPFrBfr7KOFagycvS5CJIAKBVI0bQ6gyNCtaa5HiF+Ptq27ECTVqapsJzBAkAtFbECFqlIZHVQdLR31dfHC/QpKWpKiwlSACgNSJG0GoN7hastTNGq1MHu3ZkFerupakqKK0weywAgJeIEbRqV0UEae2MeHXuYFfmiUJNXEKQAEBrQ4yg1RsYHqSXZ45WaIBdX550acLiVJ0tIUgAoLUgRtAm9A8L1MszRis0wKFd2S6NX5yiMwQJALQKxAjajH5hgVo3M15dAh3ak1OkCYtTdLq43OyxAACXQYygTenbNVDrZo5W1/NBMn5xivIJEgBo0YgRtDl9ugRo3czRCgtyaF9uscYvStGpIoIEAFoqYgRtUu8uAVo3M0HhQU7tzyvWuEWblecqM3ssAEAdGhQjCxYsUM+ePeV0OhUfH6+0tLSLrnvDDTfIYrFccLv11lsbPDRQH71CO2jdzNGKCHbq4KkSjVucolyCBABaHK9jZP369Zo1a5bmzp2rjIwMxcTEaMyYMcrLy6tz/ddff13Z2dk1t507d8pms+mnP/3pFQ8PXE7P0A5aPzNBkSF+OnSqROMWpSinkCABgJbE6xiZP3++ZsyYoaSkJA0aNEgvvvii/P39tWzZsjrX79Spk8LDw2tu77//vvz9/YkRNJvunf21buZoRYb46XB+icYt2qzswnNmjwUAOM+rGKmoqFB6eroSExO/vgOrVYmJidq8eXO97mPp0qUaN26cOnTocNF1ysvL5XK5at2AKxHdqTpIojr66cjpUo1blKKTBQQJALQEXsVIfn6+3G63wsLCai0PCwtTTk7OZbdPS0vTzp07lZycfMn15s2bp+Dg4JpbdHS0N2MCdfoqSKI7+eno+SA5QZAAgOma9dc0S5cu1dChQxUXF3fJ9ebMmaPCwsKa2/Hjx5tpQrR1UR39tX5mgrp38texM6Uau3Czjp8pNXssAGjXvIqR0NBQ2Ww25ebm1lqem5ur8PDwS25bUlKidevWafr06Zd9HIfDoaCgoFo3oLF0C/HT+ntGq2dnf2WdPadxi1IIEgAwkVcxYrfbFRsbq40bN9Ys83g82rhxoxISEi657SuvvKLy8nLdfffdDZsUaEQRwX5aNzNBvUI76ERBdZAcO02QAIAZvD5MM2vWLC1evFgrV67U7t27dd9996mkpERJSUmSpMmTJ2vOnDkXbLd06VLdcccd6ty585VPDTSC8GCn1s0crd5dvgqSzTp6usTssQCg3fHxdoOxY8fq1KlTeuKJJ5STk6Phw4drw4YNNV9qPXbsmKzW2o2zd+9ebdq0Se+9917jTA00krAgp9bNGK3xi1OqT4y2KEUvzxitnqEX/7UXAKBxWQzDMMwe4nJcLpeCg4NVWFjI90fQJPKKyjRxcar25xUrLMhRcwgHANBw9X3/5to0gKSugU6tnTFa/cMClOsq19iFm3XwVLHZYwFAu0CMAOd1CXRo7YzRGhgeqLyico1blKIDeQQJADQ1YgT4htAAh9Ykx2tgeKBOnQ+S/blFZo8FAG0aMQJ8S+eA6k9IBkUEKb+4XOMXp2gfQQIATYYYAerQqYNda5LjNbhbkPKLKzR+UYr25HCNJABoCsQIcBEdzwfJ0MhgnS6p0ITFqdqdTZAAQGMjRoBLCPG366Xp8RoWFawzJRWasDhFX54sNHssAGhTiBHgMoL9fbV6erxiokN0trRSE5ekaucJggQAGgsxAtRDsJ+vVk+P0/DoEBWcD5LMLIIEABoDMQLUU5CzOkiu7h6iwnOVmrgkRTuyCsweCwBaPWIE8EKg01crp8UptkdHucqqNHFJqrYfLzB7LABo1YgRwEtfBcmonh1VVFalSUtSte3YWbPHAoBWixgBGiDA4aMVSXGK69VJReVVmrQ0TelHCRIAaAhiBGigDg4frUgapdG9O6m4vEqTl6Zq65EzZo8FAK0OMQJcAX+7j5ZNHaWE3p1VUuHWlGVp2kKQAIBXiBHgCn0VJNf2/TpIUg+dNnssAGg1iBGgEfjZbVo6ZZSu6xeq0gq3pi7fohSCBADqhRgBGonT16bFk0fqu/276FylW9NXbFEGv7IBgMsiRoBG5PS1adGkWH2nb2jNIRtOHQ8Al0aMAI3M6WvTosmxNechmbwsTftyi8weCwBaLGIEaAJffak15vzVficuSdXh/BKzxwKAFokYAZrIV2dqHRgeqFNF5Zq4OEVZZ0vNHgsAWhxiBGhCIf52vZQcr95dOuhkYZkmLE5VrqvM7LEAoEUhRoAmFhrg0Nrk0ereyV/HzpRqwuIU5ReXmz0WALQYxAjQDMKDnVqTHK+IYKcOnirRpKVpKiitMHssAGgRiBGgmUR38tea5HiFBji0O9ulKcu3qKis0uyxAMB0xAjQjHp3CdCa5Hh19PfVF8cLNH3FVp2rcJs9FgCYihgBmtmA8ECtnh6vQKeP0o6c0czVW1VWSZAAaL+IEcAEQyKDtSIpTv52mz7dn68H1mao0u0xeywAMAUxApgktkdHLZkyUg4fqz7YnaeH12+X22OYPRYANDtiBDDRNX1CtXBSrHxtFr29I1uPvLpDHoIEQDtDjAAmu2FAV/1l/NWyWS16LSNLj7+1U4ZBkABoP4gRoAW4eUi45t8VI4tFWpN6TL99ezdBAqDdIEaAFuL24ZH63Y+HSpKWbDqsZ9/fZ/JEANA8iBGgBRk7qrv+57ZBkqQ/f3hAf/v4oMkTAUDTI0aAFmbqtb30q5sHSpJ+v2GPVnx22OSJAKBpESNAC3TfDX30sx/0kyT9zz92af2WYyZPBABNhxgBWqifJ/bTjOt6SZJmv56pt7afMHkiAGgaxAjQQlksFv36lqt09+juMgxp1t+/0IadOWaPBQCNjhgBWjCLxaKnfjhEP4mNkttj6MGXM/TR3jyzxwKARkWMAC2c1WrR7+8cpluHRajSbeje1en6/GC+2WMBQKMhRoBWwGa16Lmxw5V4VZjKqzxKXrlV6UfPmD0WADQKYgRoJXxtVr0wYYSu6xeq0gq3pi7bosysQrPHAoArRowArYjT16ZFk0YqrlcnFZVXadKyVO3NKTJ7LAC4IsQI0Mr42W1aNnWUYqJDVFBaqYlLUnXoVLHZYwFAgxEjQCsU4PDRqqQ4DYoIUn5xuSYuSdXxM6VmjwUADUKMAK1UsL+vVk+PU9+uAcouLNOEJSnKKSwzeywA8BoxArRinQMcWpMcrx6d/XX8zDlNWJKiU0XlZo8FAF4hRoBWLizIqTXJ8eoW7NShUyWatDRVBaUVZo8FAPVGjABtQFRHf62dMVpdAh3ak1OkycvS5CqrNHssAKiXBsXIggUL1LNnTzmdTsXHxystLe2S6xcUFOj+++9XRESEHA6H+vfvr3feeadBAwOoW8/QDlqbHK9OHezakVWoacu3qLSiyuyxAOCyvI6R9evXa9asWZo7d64yMjIUExOjMWPGKC+v7utlVFRU6MYbb9SRI0f06quvau/evVq8eLEiIyOveHgAtfULC9SqaXEKcvpo69GzmrFqq8oq3WaPBQCXZDEMw/Bmg/j4eI0aNUovvPCCJMnj8Sg6OloPPvigZs+efcH6L774ov7whz9oz5498vX1bdCQLpdLwcHBKiwsVFBQUIPuA2hPMo6d1aQlqSqpcOv7A7vqxbtjZffhqCyA5lXf92+vXp0qKiqUnp6uxMTEr+/AalViYqI2b95c5zb/93//p4SEBN1///0KCwvTkCFD9Mwzz8jtvvj/WysvL5fL5ap1A1B/V3fvqKVTR8npa9WHe/L08PptqnJ7zB4LAOrkVYzk5+fL7XYrLCys1vKwsDDl5OTUuc2hQ4f06quvyu1265133tHjjz+uP/3pT/rNb35z0ceZN2+egoODa27R0dHejAlA0ujenbVw0kjZbVa9k5mjR17dIY/Hqw9CAaBZNPnnth6PR127dtWiRYsUGxursWPH6tFHH9WLL7540W3mzJmjwsLCmtvx48ebekygTbq+fxe9MGGEbFaLXt92Qo++uVNeHpkFgCbnVYyEhobKZrMpNze31vLc3FyFh4fXuU1ERIT69+8vm81Ws+yqq65STk6OKirqPheCw+FQUFBQrRuAhrlpcLieHTtcFov0ctoxPf3P3QQJgBbFqxix2+2KjY3Vxo0ba5Z5PB5t3LhRCQkJdW5z7bXX6sCBA/J4vj5evW/fPkVERMhutzdwbADe+GFMN/3+zmGSpGWfHdaf3ttn8kQA8DWvD9PMmjVLixcv1sqVK7V7927dd999KikpUVJSkiRp8uTJmjNnTs369913n86cOaOHHnpI+/bt09tvv61nnnlG999/f+P9FQAu666R0Xr69sGSpBc+OqAFHx0weSIAqObj7QZjx47VqVOn9MQTTygnJ0fDhw/Xhg0bar7UeuzYMVmtXzdOdHS03n33Xf385z/XsGHDFBkZqYceeki/+tWvGu+vAFAvkxJ66lylW8+8s0d/eHevnL42Tf9OL7PHAtDOeX2eETNwnhGgcT33wT4998F+SdIzPxqqCfHdTZ4IQFvUJOcZAdA2PPSDfrrn+t6SpEffzNQb27JMnghAe0aMAO2QxWLR7JsHakpCDxmG9Iu/f6F3MrPNHgtAO0WMAO2UxWLR3NsG666RUfIY0s9e3qYP9+RefkMAaGTECNCOWa0WzfvxMP0wppuqPIbufSlDnx3IN3ssAO0MMQK0czarRX+6K0Y3DgpTRZVHySu3asuRM2aPBaAdIUYAyNdm1QsTRui7/bvoXKVbScu36IvjBWaPBaCdIEYASJIcPjYtvDtW8b06qbi8SpOXpWl3NlfMBtD0iBEANfzsNi2dOkojuoeo8FylJi1N1YG8YrPHAtDGESMAaglw+GhFUpwGdwtSfnGFJi5J0bHTpWaPBaANI0YAXCDYz1erp8erX9cA5brKNWFJik4WnDN7LABtFDECoE6dOti1JjlePTv7K+vsOU1ckqq8ojKzxwLQBhEjAC6qa5BTa2aMVmSInw7nl2jSkjSdLakweywAbQwxAuCSIkP8tHZGvMKCHNqbW6RJy1JVeK7S7LEAtCHECIDL6tG5g9Ykx6tzB7t2nnApaXmaSsqrzB4LQBtBjACol75dA7V6eryCnD7KOFag5JVbVVbpNnssAG0AMQKg3gZ1C9Kq6fEKcPho86HTuveldJVXESQArgwxAsArw6NDtGzqKDl9rfp47yn97OVtqnJ7zB4LQCtGjADwWlyvTlo8eaTsNqve/TJXv3jlC7k9htljAWiliBEADXJdvy7668Sr5WO16K3tJ/XoG5nyECQAGoAYAdBgiYPC9Py4EbJapHVbjuupf+6SYRAkALxDjAC4IrcOi9AffhIjSVrx+RH977t7CRIAXiFGAFyxO2Oj9Js7hkiS/vbxQb3w4QGTJwLQmhAjABrF3aN76LFbr5Ik/en9fVry6SGTJwLQWhAjABpN8nW99Ysb+0uSfvP2bq1OOWryRABaA2IEQKN64Pt9dd8NfSRJj7+5U6+mZ5k8EYCWjhgB0KgsFoseGTNAU6/pKUl65NUv9M8dJ80dCkCLRowAaHQWi0VzbxukcaOi5TGkh9dt1we7cs0eC0ALRYwAaBIWi0W//dFQ3TG8m6o8hv5rTYY+3X/K7LEAtEDECIAmY7Na9MefxujmweGqcHs0Y9VWpR46bfZYAFoYYgRAk/KxWfXn8SP0vQFdVFbp0bQVW7T9eIHZYwFoQYgRAE3O7mPV3+6O1TV9Oqukwq3JS1P15clCs8cC0EIQIwCahdPXpsWTRyq2R0e5yqo0aWmaDuQVmT0WgBaAGAHQbDo4fLQ8aZSGRgbrTEmFJixO1ZH8ErPHAmAyYgRAswpy+mrVtDgNCAtUXlG5Ji5J1YmCc2aPBcBExAiAZtexg12rk+PUO7SDThSc08TFKcpzlZk9FgCTECMATNE10Kk1M+IV1dFPR06XauKSVJ0uLjd7LAAmIEYAmCYi2E9rk0crPMip/XnFmrQ0TYWllWaPBaCZESMATNW9s7/WzIhXaIBdu7JdmrI8TcXlVWaPBaAZESMATNenS4BWT49XiL+vth8v0PQVW3Suwm32WACaCTECoEW4KiJIq6bFKdDho9TDZ3TPS+kqryJIgPaAGAHQYgyLCtHypFHy87Xp3/tO6YG121Tp9pg9FoAmRowAaFFG9uykJVNGyu5j1fu7cjXr71/I7THMHgtAEyJGALQ41/YN1Yt3Xy1fm0X/+OKkZr+2Qx6CBGiziBEALdL3B4bpz+NGyGqRXknP0v/840sZBkECtEXECIAW6z+GRuhPd8XIYpFWbT6q3/1rD0ECtEHECIAW7UcjovTbO4ZKkhb++5Ce37jf5IkANDZiBECLNyG+u574z0GSpOc+2K+Fnxw0eSIAjYkYAdAqTPtOL/1yzABJ0rx/7dGqzUfMHQhAoyFGALQa93+vrx74Xl9J0hNvfam/bz1u8kQAGgMxAqBV+cVN/TXt2l6SpF+9tkP/98VJkycCcKUaFCMLFixQz5495XQ6FR8fr7S0tIuuu2LFClksllo3p9PZ4IEBtG8Wi0WP/+dVmhDfXYYh/Xz9dr37ZY7ZYwG4Al7HyPr16zVr1izNnTtXGRkZiomJ0ZgxY5SXl3fRbYKCgpSdnV1zO3r06BUNDaB9s1gs+s3tQ/TjEZFyeww9uHabPtl3yuyxADSQ1zEyf/58zZgxQ0lJSRo0aJBefPFF+fv7a9myZRfdxmKxKDw8vOYWFhZ2RUMDgNVq0f/+ZJhuGRquCrdHM1dt1eaDp80eC0ADeBUjFRUVSk9PV2Ji4td3YLUqMTFRmzdvvuh2xcXF6tGjh6Kjo3X77bfryy+/vOTjlJeXy+Vy1boBwLf52Kx6buwIfX9gV5VXeTR95RZlHDtr9lgAvORVjOTn58vtdl/wyUZYWJhycuo+ZjtgwAAtW7ZMb731ll566SV5PB5dc801ysrKuujjzJs3T8HBwTW36Ohob8YE0I7Yfaz668Sr9Z2+oSqtcGvKsjTtPFFo9lgAvNDkv6ZJSEjQ5MmTNXz4cF1//fV6/fXX1aVLFy1cuPCi28yZM0eFhYU1t+PH+fkegItz+tq0aHKsRvXsqKKyKk1amqp9uUVmjwWgnryKkdDQUNlsNuXm5tZanpubq/Dw8Hrdh6+vr0aMGKEDBw5cdB2Hw6GgoKBaNwC4FH+7j5ZNHaVhUcE6W1qpiUtSdTi/xOyxANSDVzFit9sVGxurjRs31izzeDzauHGjEhIS6nUfbrdbmZmZioiI8G5SALiMQKevVk2L08DwQJ0qKtfExSnKOltq9lgALsPrwzSzZs3S4sWLtXLlSu3evVv33XefSkpKlJSUJEmaPHmy5syZU7P+U089pffee0+HDh1SRkaG7r77bh09elTJycmN91cAwHkh/na9lByv3l066GRhmSYsTlWuq8zssQBcgo+3G4wdO1anTp3SE088oZycHA0fPlwbNmyo+VLrsWPHZLV+3Thnz57VjBkzlJOTo44dOyo2Nlaff/65Bg0a1Hh/BQB8Q2iAQ2uTR+uuhZt17EypJixO0fp7EhQa4DB7NAB1sBiGYZg9xOW4XC4FBwersLCQ748AqLfjZ0p118LNyi4s01URQXp5RrxC/O1mjwW0G/V9/+baNADarOhO/lqTHK/QAId2Z7s0ZfkWFZVVmj0WgG8hRgC0ab27BGhNcrw6+vvqi+MFmr5iq0orqsweC8A3ECMA2rwB4YFaPT1egU4fpR05o3tWp6us0m32WADOI0YAtAtDIoO1IilO/nabPt2frwfWZqjS7TF7LAAiRgC0I7E9OmrJlJFy+Fj1we48Pbxuu6oIEsB0xAiAduWaPqF6cVKsfG0WvZ2ZrUde2yGPp8X/qBBo04gRAO3O9wZ01V/GXy2b1aLXM07o8bd2qhWc5QBos4gRAO3SzUPCNf+uGFks0prUY/rt27sJEsAkxAiAduv24ZH63Y+HSpKWbDqsZ9/fZ/JEQPtEjABo18aO6q7/ua368hR//vCA/vrxxa8oDqBpECMA2r2p1/bSr24eKEn63w17tfyzwyZPBLQvxAgASLrvhj762Q/6SZKe/McurUs7ZvJEQPtBjADAeT9P7KcZ1/WSJM15I1Nvbjth8kRA+0CMAMB5FotFv77lKt09ursMQ/rFK19ow85ss8cC2jxiBAC+wWKx6KkfDtGdV0fJ7TH04Mvb9NGePLPHAto0YgQAvsVqteh/fzJMtw6LUKXb0L0vpevzg/lmjwW0WcQIANTBZrXoubHDlXhVmMqrPEpeuVXpR8+YPRbQJhEjAHARvjarXpgwQtf1C1VphVtTl21RZlah2WMBbQ4xAgCX4PS1adGkkYrr1UlF5VWatCxVe3OKzB4LaFOIEQC4DD+7TcumjlJMdIgKSis1cUmqDp0qNnssoM0gRgCgHgIcPlqVFKdBEUHKLy7XxCWpOn6m1OyxgDaBGAGAegr299Xq6XHq2zVA2YVlmrAkRdmF58weC2j1iBEA8ELnAIfWJMerR2d/HT9zThOXpOpUUbnZYwGtGjECAF4KC3JqTXK8ugU7dehUiSYtTdXZkgqzxwJaLWIEABogqqO/1s4YrS6BDu3JKdKU5WlylVWaPRbQKhEjANBAPUM7aG1yvDp1sGtHVqGmLd+i0ooqs8cCWh1iBACuQL+wQK2aFqcgp4+2Hj2rGau2qqzSbfZYQKtCjADAFRoSGawV0+LUwW7TZwdO67/WZKiiymP2WECrQYwAQCO4untHLZ06Sk5fqz7ck6eH1m1TlZsgAeqDGAGARjK6d2ctnDRSdptV/9qZo1++ukMej2H2WECLR4wAQCO6vn8XvTBhhGxWi97YdkKPvrlThkGQAJdCjABAI7tpcLieHTtcFov0ctoxPfXPXQQJcAnECAA0gR/GdNPv7xwmSVr+2RH98b29Jk8EtFzECAA0kbtGRuvp2wdLkhZ8dFALPjpg8kRAy0SMAEATmpTQU7++ZaAk6Q/v7tXSTYdNnghoeYgRAGhiM7/bRw8n9pMkPf3PXVqbeszkiYCWhRgBgGbw0A/66Z7re0uSHn0zU69nZJk8EdByECMA0AwsFotm3zxQUxJ6yDCk/37lC72TmW32WECLQIwAQDOxWCyae9tg/TQ2Sh5D+tnL2/ThnlyzxwJMR4wAQDOyWi363Z3DdFtMN1V5DN37UoY27c83eyzAVMQIADQzm9Wi+XfF6MZBYaqo8mjGqq3acuSM2WMBpiFGAMAEvjarXpgwQt/t30XnKt1KWr5FXxwvMHsswBTECACYxOFj08K7YxXfq5OKy6s0eVmadme7zB4LaHbECACYyM9u09KpozSie4gKz1Vq0tJUHcgrNnssoFkRIwBgsgCHj1YkxWlwtyDlF1do4pIUHTtdavZYQLMhRgCgBQj289Xq6fHq1zVAua5yTViSopMF58weC2gWxAgAtBCdOti1JjlePTv7K+vsOU1ckqq8ojKzxwKaHDECAC1I1yCn1swYrcgQPx3OL9GkJWk6U1Jh9lhAkyJGAKCFiQzx09oZ8eoa6NDe3CJNXpaqwnOVZo8FNBliBABaoB6dO2jtjHh17mDXzhMuJS1PU0l5ldljAU2iQTGyYMEC9ezZU06nU/Hx8UpLS6vXduvWrZPFYtEdd9zRkIcFgHalb9dArZ4eryCnjzKOFSh55VaVVbrNHgtodF7HyPr16zVr1izNnTtXGRkZiomJ0ZgxY5SXl3fJ7Y4cOaL//u//1nXXXdfgYQGgvRnULUirpscrwOGjzYdO68ZnP9Gs9du1/LPDSj96RucqiBO0fhbDMAxvNoiPj9eoUaP0wgsvSJI8Ho+io6P14IMPavbs2XVu43a79d3vflfTpk3Tp59+qoKCAr355pv1fkyXy6Xg4GAVFhYqKCjIm3EBoE1IO3xG01ZsUfG3DtVYLVK/roEaGhWsYVHBGhoZrKsiguT0tZk0KfC1+r5/+3hzpxUVFUpPT9ecOXNqllmtViUmJmrz5s0X3e6pp55S165dNX36dH366aeXfZzy8nKVl5fX/LfLxemRAbRvcb06adOvvqf0o2eVeaJQmVmF+iKrUPnF5dqbW6S9uUV6NT1LUvWF+PqHBWpYZLCGng+UgRGBcvgQKGiZvIqR/Px8ud1uhYWF1VoeFhamPXv21LnNpk2btHTpUm3fvr3ejzNv3jw9+eST3owGAG1eiL9dP7gqTD+4qvo12DAM5brKz8dJgXacj5TTJRXane3S7myX1m89LknytVk0IDxQQyODNTQyRMOigtU/LFB2H37HAPN5FSPeKioq0qRJk7R48WKFhobWe7s5c+Zo1qxZNf/tcrkUHR3dFCMCQKtlsVgUHuxUeLBTNw76OlCyC8u0I6tQmScKtCOrUDtPFOpsaaV2nnBp5wmXXlZ1oNhtVg2MqA6U6kM8IeoXFiBfG4GC5uVVjISGhspmsyk3N7fW8tzcXIWHh1+w/sGDB3XkyBHddtttNcs8Hk/1A/v4aO/everTp88F2zkcDjkcDm9GAwCoOlC6hfipW4ifbh5S/bpsGIayzp6r/gTl/KcnO7IK5Cqr0o6sQu3IKtSa1Ort7T5WDYoIqvn+ydCoYPXtEiAfAgVNyKsYsdvtio2N1caNG2t+nuvxeLRx40Y98MADF6w/cOBAZWZm1lr22GOPqaioSM8//zyfdgBAM7BYLIru5K/oTv66ZWiEpOpAOXam9BtxUv0JSlF5lbYfL9D24wU12zt9rRrc7XycnP8UpXeXANmsFpP+IrQ1Xh+mmTVrlqZMmaKRI0cqLi5Ozz33nEpKSpSUlCRJmjx5siIjIzVv3jw5nU4NGTKk1vYhISGSdMFyAEDzsVgs6tG5g3p07qD/HNZNkuTxGDp6plQ7sgqUmVX9KcrOE4UqqXAr/ehZpR89W7O9v92mwd2Car5/MjQqWL06d5CVQEEDeB0jY8eO1alTp/TEE08oJydHw4cP14YNG2q+1Hrs2DFZrXycBwCtjdVqUa/QDuoV2kG3D4+UVB0oh/JLlHmiQJlZLmWeKNDOEy6VVri15chZbTnydaAEOHw0uFv1IZ4hkcEaFhWiHp38CRRcltfnGTED5xkBgJbD7TF08FRxzacnO7IK9OVJl8qrPBesG+j0qTm8MzQqWMMiQxTdyU8WC4HSHtT3/ZsYAQBcsSq3RwdOFVf/iud8pOzKdqmijkAJ9vP9RpxU/29kCIHSFhEjAABTVbo92pdbVBMnmScKtTvbpUr3hW87Hf19NTQqpNaJ2iKCnQRKK0eMAABanPIqt/blFJ+Pk+rzoOzNKVKV58K3otAA+zcO8VR/UTYsyGnC1GgoYgQA0CqUVbq1N6fo/BlkC5R5wqV9uUVy1xEoXQMdXx/iOf9F2a6BBEpLRYwAAFqtskq3dmW7vj7Ek1Wo/XlFqqNPFB7krPX9k6GRweocwIkzWwJiBADQppRWVGl3tqvmS7I7ThTq4Kli1fUuFhniV/MJyleHejp2sDf/0O0cMQIAaPOKy6u066Sr+kRt578ke+hUSZ3rRnfyq3WhwCHdghXs79vME7cvxAgAoF1ylVXqyxPVJ2jLPOFSZlaBjpwurXPdHp39a10ocEhkkAKdBEpjIUYAADivsLRSO09+40KBJwp0/My5OtftHdqh1uGdwZHBCnA06UXu2yxiBACASzhbUqGdJwtrnajtRMGFgWKxSH26BNS6UOCgbkHytxMol0OMAADgpdPF5TUXCNxxPlCyC8suWM9qkfp2Dah1ocBBEUFy+tpMmLrlIkYAAGgEp4rKvxEn1Sdqyysqv2A9m9Wifl0Dzn//pPpEbQPDA9t1oBAjAAA0kVxXWc3Pi3eev1hgfnHFBev5WC0aEB5Y60KBA8IDZfdpH1e3J0YAAGgmhmEox1WmHVm1D/GcKbkwUOw2a3WgnD9R25DIYA0ID5Svre0FCjECAICJDMPQiYJzteIk80ShCkorL1jX7mPVVRFBtc4i269rgHxaeaAQIwAAtDCGYSjr7DntOP/z4q9Cpais6oJ1nb5WDYoIqnWhwD5dAmSztp4rGRMjAAC0Ah6PoWNnSmt9/2TnCZeKyy8MFD9fmwZ3C6q5UODQyBD1Du0gawsNFGIEAIBWyuMxdPh0ydeHeLIKtfNkoUor3Bes28Fu0+DI2hcK7Nm5ZQQKMQIAQBvi9hg6nF9cfYjn/Bdld54sVFml54J1Ax0+GvKNOBkWFazunfxlsTRvoBAjAAC0cVVujw6eKjl/aKf6p8a7TrpUXnVhoAQ5fc7HSUjNuVCiOvo1aaAQIwAAtEOVbo/25xafj5MCZWYVand2kSrcFwZKiL9vzScnP42NVs/QDo06CzECAAAkSRVVHu3LLVLmia8P8ezJcanS/XUCvHJvgkb17NSoj1vf92+u8gMAQBtn97FqyPkTrI2Pq15WXuXW3pzzgXK8UIMizPs/+8QIAADtkMPHpmFRIRoWFaKJ8ebO0rpP7QYAAFo9YgQAAJiKGAEAAKYiRgAAgKmIEQAAYCpiBAAAmIoYAQAApiJGAACAqYgRAABgKmIEAACYihgBAACmIkYAAICpiBEAAGCqVnHVXsMwJEkul8vkSQAAQH199b791fv4xbSKGCkqKpIkRUdHmzwJAADwVlFRkYKDgy/67xbjcrnSAng8Hp08eVKBgYGyWCyNdr8ul0vR0dE6fvy4goKCGu1+2yL2lXfYX/XHvqo/9lX9sa/qryn3lWEYKioqUrdu3WS1XvybIa3ikxGr1aqoqKgmu/+goCCerPXEvvIO+6v+2Ff1x76qP/ZV/TXVvrrUJyJf4QusAADAVMQIAAAwVbuOEYfDoblz58rhcJg9SovHvvIO+6v+2Ff1x76qP/ZV/bWEfdUqvsAKAADarnb9yQgAADAfMQIAAExFjAAAAFMRIwAAwFRtPkYWLFignj17yul0Kj4+XmlpaZdc/5VXXtHAgQPldDo1dOhQvfPOO800qfm82VcrVqyQxWKpdXM6nc04rXn+/e9/67bbblO3bt1ksVj05ptvXnabjz/+WFdffbUcDof69u2rFStWNPmcLYG3++rjjz++4HllsViUk5PTPAObaN68eRo1apQCAwPVtWtX3XHHHdq7d+9lt2uPr1kN2Vft9TXrb3/7m4YNG1ZzQrOEhAT961//uuQ2Zjyn2nSMrF+/XrNmzdLcuXOVkZGhmJgYjRkzRnl5eXWu//nnn2v8+PGaPn26tm3bpjvuuEN33HGHdu7c2cyTNz9v95VUfba+7OzsmtvRo0ebcWLzlJSUKCYmRgsWLKjX+ocPH9att96q733ve9q+fbsefvhhJScn6913323iSc3n7b76yt69e2s9t7p27dpEE7Ycn3zyie6//36lpKTo/fffV2VlpW666SaVlJRcdJv2+prVkH0ltc/XrKioKP3ud79Tenq6tm7dqu9///u6/fbb9eWXX9a5vmnPKaMNi4uLM+6///6a/3a73Ua3bt2MefPm1bn+XXfdZdx66621lsXHxxv33HNPk87ZEni7r5YvX24EBwc303QtlyTjjTfeuOQ6jzzyiDF48OBay8aOHWuMGTOmCSdreeqzrz766CNDknH27Nlmmakly8vLMyQZn3zyyUXXac+vWd9Un33Fa9bXOnbsaCxZsqTOfzPrOdVmPxmpqKhQenq6EhMTa5ZZrVYlJiZq8+bNdW6zefPmWutL0pgxYy66flvRkH0lScXFxerRo4eio6MvWdrtXXt9Xl2J4cOHKyIiQjfeeKM+++wzs8cxRWFhoSSpU6dOF12H51a1+uwridcst9utdevWqaSkRAkJCXWuY9Zzqs3GSH5+vtxut8LCwmotDwsLu+jx55ycHK/Wbysasq8GDBigZcuW6a233tJLL70kj8eja665RllZWc0xcqtyseeVy+XSuXPnTJqqZYqIiNCLL76o1157Ta+99pqio6N1ww03KCMjw+zRmpXH49HDDz+sa6+9VkOGDLnoeu31Neub6ruv2vNrVmZmpgICAuRwOHTvvffqjTfe0KBBg+pc16znVKu4ai9anoSEhFplfc011+iqq67SwoUL9fTTT5s4GVqzAQMGaMCAATX/fc011+jgwYN69tlntXr1ahMna17333+/du7cqU2bNpk9SotX333Vnl+zBgwYoO3bt6uwsFCvvvqqpkyZok8++eSiQWKGNvvJSGhoqGw2m3Jzc2stz83NVXh4eJ3bhIeHe7V+W9GQffVtvr6+GjFihA4cONAUI7ZqF3teBQUFyc/Pz6SpWo+4uLh29bx64IEH9M9//lMfffSRoqKiLrlue33N+oo3++rb2tNrlt1uV9++fRUbG6t58+YpJiZGzz//fJ3rmvWcarMxYrfbFRsbq40bN9Ys83g82rhx40WPlSUkJNRaX5Lef//9i67fVjRkX32b2+1WZmamIiIimmrMVqu9Pq8ay/bt29vF88owDD3wwAN644039OGHH6pXr16X3aa9Prcasq++rT2/Znk8HpWXl9f5b6Y9p5r067EmW7duneFwOIwVK1YYu3btMmbOnGmEhIQYOTk5hmEYxqRJk4zZs2fXrP/ZZ58ZPj4+xh//+Edj9+7dxty5cw1fX18jMzPTrD+h2Xi7r5588knj3XffNQ4ePGikp6cb48aNM5xOp/Hll1+a9Sc0m6KiImPbtm3Gtm3bDEnG/PnzjW3bthlHjx41DMMwZs+ebUyaNKlm/UOHDhn+/v7GL3/5S2P37t3GggULDJvNZmzYsMGsP6HZeLuvnn32WePNN9809u/fb2RmZhoPPfSQYbVajQ8++MCsP6HZ3HfffUZwcLDx8ccfG9nZ2TW30tLSmnV4zarWkH3VXl+zZs+ebXzyySfG4cOHjR07dhizZ882LBaL8d577xmG0XKeU206RgzDMP7yl78Y3bt3N+x2uxEXF2ekpKTU/Nv1119vTJkypdb6f//7343+/fsbdrvdGDx4sPH2228388Tm8WZfPfzwwzXrhoWFGbfccouRkZFhwtTN76ufn3779tX+mTJlinH99ddfsM3w4cMNu91u9O7d21i+fHmzz20Gb/fV73//e6NPnz6G0+k0OnXqZNxwww3Ghx9+aM7wzayu/SSp1nOF16xqDdlX7fU1a9q0aUaPHj0Mu91udOnSxfjBD35QEyKG0XKeUxbDMIym/ewFAADg4trsd0YAAEDrQIwAAABTESMAAMBUxAgAADAVMQIAAExFjAAAAFMRIwAAwFTECAAAMBUxAgAATEWMAAAAUxEjAADAVMQIAAAw1f8DnvAOePzft1wAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5222765656906478\n"
     ]
    }
   ],
   "source": [
    "def import_data_set():\n",
    "    # Grab dataset from file\n",
    "    car_dataframe = pd.read_csv(\"data/car details v4.csv\")\n",
    "    data = (car_dataframe.loc[:, [\"Fuel Tank Capacity\", \"Year\", \"Price\"]]).to_numpy().astype('float64')\n",
    "    return data\n",
    "\n",
    "def create_data_sets(data):\n",
    "    # Use simple data cleaning of dropping all null data\n",
    "    data_clean = data[~np.isnan(data).any(axis=1)]\n",
    "\n",
    "    # Suffle data insure equal distabution between train, validation, and test sets (Also use seed to insure determinisum)\n",
    "    rng = np.random.default_rng(seed=42)\n",
    "    rng.shuffle(data_clean)\n",
    "\n",
    "    # Split up data set into train, validation, and test\n",
    "    train_set_end_index = int(data_clean.shape[0]/2)\n",
    "    validation_set_end_index = int(3*data_clean.shape[0]/4)\n",
    "    train_set = data_clean[:train_set_end_index]\n",
    "    validation_set = data_clean[train_set_end_index:validation_set_end_index]\n",
    "    test_set = data_clean[validation_set_end_index:]\n",
    "\n",
    "    train_set_inputs = train_set[:, :2]\n",
    "    train_set_targets = train_set[:, 2:]\n",
    "    validation_set_inputs = validation_set[:, :2]\n",
    "    validation_set_targets = validation_set[:, 2:]\n",
    "    test_set_inputs = test_set[:, :2]\n",
    "    test_set_targets = test_set[:, 2:]\n",
    "\n",
    "    return train_set_inputs, train_set_targets, validation_set_inputs, validation_set_targets, test_set_inputs, test_set_targets\n",
    "\n",
    "def normalize_datasets(train_set_inputs, train_set_targets, validation_set_inputs, validation_set_targets, test_set_inputs, test_set_targets):\n",
    "    normalized_train_set_inputs = train_set_inputs\n",
    "    normalized_validation_set_inputs = validation_set_inputs\n",
    "    normalized_test_set_inputs = test_set_inputs\n",
    "\n",
    "    for j in range(train_set_inputs.shape[1]):\n",
    "        train_set_column = normalized_train_set_inputs[:, j:j+1]\n",
    "        validation_set_column = normalized_validation_set_inputs[:, j:j+1]\n",
    "        test_set_column = normalized_test_set_inputs[:, j:j+1]\n",
    "\n",
    "        sigma = train_set_column.std()\n",
    "        mean = train_set_column.mean()\n",
    "\n",
    "        normalized_train_set_inputs[:, j:j+1] = (train_set_column - mean) / sigma / 100\n",
    "        normalized_validation_set_inputs[:, j:j+1] = (validation_set_column - mean) / sigma / 100\n",
    "        normalized_test_set_inputs[:, j:j+1] = (test_set_column - mean) / sigma / 100\n",
    "        \n",
    "    train_set_targets_mean = train_set_targets.mean()\n",
    "\n",
    "    normalized_train_set_targets = train_set_targets / train_set_targets_mean\n",
    "    normalized_validation_set_targets = validation_set_targets / train_set_targets_mean\n",
    "    normalized_test_set_targets = test_set_targets / train_set_targets_mean\n",
    "\n",
    "    return normalized_train_set_inputs, normalized_train_set_targets, normalized_validation_set_inputs, normalized_validation_set_targets, normalized_test_set_inputs, normalized_test_set_targets \n",
    "\n",
    "def get_next_polynomial_set(normalized_trained_inputs):\n",
    "    num_of_examples = normalized_trained_inputs.shape[0]\n",
    "    num_of_inputs = normalized_trained_inputs.shape[1]\n",
    "\n",
    "    next_polynomal_set = np.empty([num_of_examples, num_of_inputs**2+num_of_inputs])\n",
    "\n",
    "    for i in range(num_of_examples):\n",
    "        for j in range(num_of_inputs):\n",
    "            next_polynomal_set[i, j] = normalized_trained_inputs[i, j]\n",
    "\n",
    "    for i in range(num_of_examples):\n",
    "        index = 0\n",
    "        for j in range(num_of_inputs):\n",
    "            for k in range(num_of_inputs):\n",
    "                next_polynomal_set[i, index+num_of_inputs] = normalized_trained_inputs[i, j] * normalized_trained_inputs[i, k]\n",
    "                index += 1\n",
    "            \n",
    "    return next_polynomal_set \n",
    "\n",
    "\n",
    "def create_polynomial_datasets(normalized_train_set_inputs, normalized_validation_set_inputs, normalized_test_set_inputs, polynomial_degree):\n",
    "    normalized_polynomial_train_set_targets = [normalized_train_set_inputs]\n",
    "    normalized_polynomial_validation_set_inputs = [normalized_validation_set_inputs]\n",
    "    normalized_polynomial_test_set_inputs = [normalized_test_set_inputs]\n",
    "\n",
    "    for _i in range(polynomial_degree-1):\n",
    "        normalized_polynomial_train_set_targets.append(get_next_polynomial_set(normalized_polynomial_train_set_targets[-1]))\n",
    "        normalized_polynomial_validation_set_inputs.append(get_next_polynomial_set(normalized_polynomial_validation_set_inputs[-1]))\n",
    "        normalized_polynomial_test_set_inputs.append(get_next_polynomial_set(normalized_polynomial_test_set_inputs[-1]))\n",
    "\n",
    "    scaling_factor = 10\n",
    "\n",
    "    for i in range(1, len(normalized_polynomial_train_set_targets)):\n",
    "        train_set_targets = normalized_polynomial_train_set_targets[i]\n",
    "        validation_set_targets = normalized_polynomial_validation_set_inputs[i]\n",
    "        test_set_targets = normalized_polynomial_test_set_inputs[i]\n",
    "        \n",
    "        num_of_targets = train_set_targets.shape[1]\n",
    "        for j in range(num_of_targets):\n",
    "            mean = train_set_targets[:, j:j+1].mean()\n",
    "            sigma = train_set_targets[:, j:j+1].std()\n",
    "\n",
    "            train_set_targets[:, j:j+1] = (train_set_targets[:, j:j+1] - mean) / sigma / scaling_factor\n",
    "            validation_set_targets[:, j:j+1] = (validation_set_targets[:, j:j+1] - mean) / sigma / scaling_factor\n",
    "            test_set_targets[:, j:j+1] = (test_set_targets[:, j:j+1] - mean) / sigma / scaling_factor\n",
    "        \n",
    "    return normalized_polynomial_train_set_targets, normalized_polynomial_validation_set_inputs, normalized_polynomial_test_set_inputs\n",
    "\n",
    "data_set = import_data_set()\n",
    "\n",
    "train_set_inputs, train_set_targets, validation_set_inputs, validation_set_targets, test_set_inputs, test_set_targets = create_data_sets(data_set)\n",
    "normalized_train_set_inputs, normalized_train_set_targets, normalized_validation_set_inputs, normalized_validation_set_targets, normalized_test_set_inputs, normalized_test_set_targets  = normalize_datasets(train_set_inputs, train_set_targets, validation_set_inputs, validation_set_targets, test_set_inputs, test_set_targets)\n",
    "\n",
    "polynomial_degree = 4\n",
    "\n",
    "normalized_polynomial_train_set_targets, normalized_polynomial_validation_set_inputs, normalized_polynomial_test_set_inputs = create_polynomial_datasets(normalized_train_set_inputs, normalized_validation_set_inputs, normalized_test_set_inputs, polynomial_degree)\n",
    "\n",
    "cost_histories = []\n",
    "learned_wb = []\n",
    "\n",
    "for i in range(polynomial_degree):\n",
    "    w, b, cost_history = run_linear_regression(normalized_polynomial_train_set_targets[i], normalized_train_set_targets)\n",
    "    learned_wb.append((w,b))\n",
    "    cost_histories.append(cost_history)\n",
    "\n",
    "for i in range(len(cost_histories)):\n",
    "    x = cost_histories[i]\n",
    "    plt.plot(x, label = f\"polynomial degee {i+1}\") \n",
    "\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "final_cost = []\n",
    "\n",
    "for i in range(len(cost_histories)):\n",
    "    final_cost.append(cost_histories[i][-1])\n",
    "\n",
    "plt.plot(final_cost) \n",
    "plt.show()\n",
    "\n",
    "validation_set_costs = []\n",
    "\n",
    "for i in range(polynomial_degree):\n",
    "    w, b = learned_wb[i]\n",
    "    cost = compute_cost(normalized_polynomial_validation_set_inputs[i], normalized_validation_set_targets, w, b)\n",
    "    validation_set_costs.append(cost)\n",
    "\n",
    "plt.plot(validation_set_costs) \n",
    "plt.show()\n",
    "\n",
    "print(compute_cost(normalized_polynomial_test_set_inputs[-1], normalized_test_set_targets, w, b))\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}