linear_regression_multiple_variables.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "6dd4b0c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tqdm import tqdm\n",
    "from sklearn import datasets\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# best to split the data into a train & test batch (75:25 or similar)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "6815f5b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(442, 10) (442,)\n",
      "[[ 0.03807591  0.05068012  0.06169621  0.02187235 -0.0442235  -0.03482076\n",
      "  -0.04340085 -0.00259226  0.01990842 -0.01764613]\n",
      " [-0.00188202 -0.04464164 -0.05147406 -0.02632783 -0.00844872 -0.01916334\n",
      "   0.07441156 -0.03949338 -0.06832974 -0.09220405]\n",
      " [ 0.08529891  0.05068012  0.04445121 -0.00567061 -0.04559945 -0.03419447\n",
      "  -0.03235593 -0.00259226  0.00286377 -0.02593034]\n",
      " [-0.08906294 -0.04464164 -0.01159501 -0.03665645  0.01219057  0.02499059\n",
      "  -0.03603757  0.03430886  0.02269202 -0.00936191]\n",
      " [ 0.00538306 -0.04464164 -0.03638469  0.02187235  0.00393485  0.01559614\n",
      "   0.00814208 -0.00259226 -0.03199144 -0.04664087]] [151.  75. 141. 206. 135.]\n",
      "10 442\n",
      "25.0 346.0 152.13348416289594\n"
     ]
    }
   ],
   "source": [
    "# using sklean diabetes dataset\n",
    "data_X, data_y = datasets.load_diabetes(return_X_y=True)\n",
    "n = 10\n",
    "m = len(data_y)\n",
    "# peaking into the dataset\n",
    "print(data_X.shape, data_y.shape)\n",
    "print(data_X[0:5], data_y[0:5])\n",
    "print(n, m)\n",
    "print(np.min(data_y), np.max(data_y), np.mean(data_y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "07853e5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to predict y\n",
    "def predict(weights, Xs):\n",
    "    return weights[0] + weights[1:].T.dot(Xs)\n",
    "\n",
    "# evaluate total cost (predictions of entire dataset)\n",
    "def cost_function(weights, X):\n",
    "    losses = []\n",
    "    for i in range(m):\n",
    "        losses.append((predict(weights, X[i])-data_y[i])**2)\n",
    "    return (1/m)*np.sum(losses)\n",
    "\n",
    "# get derivative with respect to each individual weight\n",
    "def derive_cost(weights, X):\n",
    "    H = 0.0000001\n",
    "    derivatives = np.zeros(n+1)\n",
    "    for i in range(n+1):\n",
    "        temp = np.copy(weights)\n",
    "        temp[i] = temp[i]+H\n",
    "        a = cost_function(temp, X) - cost_function(weights, X)\n",
    "        derivatives[i] = a/H\n",
    "    return derivatives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "e7824804",
   "metadata": {},
   "outputs": [],
   "source": [
    "def absolute_loss(weights, X):\n",
    "    losses = np.zeros(m)\n",
    "    for i in range(m):\n",
    "        losses[i] = abs(predict(weights, X[i])-data_y[i])\n",
    "    return losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "e870c30d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.]\n",
      "11\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 120/120 [00:20<00:00,  5.87it/s]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe93adb1a00>]"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr = 0.02\n",
    "n_epochs = 120\n",
    "\n",
    "my_weights = np.ones(n + 1)\n",
    "print(my_weights)\n",
    "\n",
    "los = []\n",
    "abslos = []\n",
    "\n",
    "print(len(derive_cost(my_weights, data_X)))\n",
    "for epoch in tqdm(range(n_epochs)):\n",
    "    \n",
    "    temp_weights = np.copy(my_weights)\n",
    "    for i in range(len(temp_weights)):\n",
    "        temp_weights[i] = temp_weights[i]-lr*derive_cost(my_weights, data_X)[i]\n",
    "        \n",
    "    los.append(cost_function(my_weights, data_X))\n",
    "    abslos.append(absolute_loss(my_weights, data_X))\n",
    "        \n",
    "    my_weights = np.copy(temp_weights)\n",
    "\n",
    "plt.plot(los)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "6da87ed1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[151.00652329   4.18581369   1.66853063  11.12642915   8.5978327\n",
      "   4.56252785   3.89890741  -5.79473196   8.35420635  10.73848692\n",
      "   7.54541582]\n",
      "prediction 0 [264] : 150.48771703437347 145.0(absolute loss = 5.487717034373475)\n",
      "prediction 1 [264] : 150.48771703437347 145.0(absolute loss = 5.487717034373475)\n",
      "prediction 2 [99] : 150.6572865667744 83.0(absolute loss = 67.65728656677439)\n",
      "prediction 3 [354] : 154.07396178504223 272.0(absolute loss = 117.92603821495777)\n",
      "prediction 4 [170] : 148.68253653659528 47.0(absolute loss = 101.68253653659528)\n",
      "28545.143784232037\n",
      "0.42079985014873955 193.1516701520314 64.5817732674933\n"
     ]
    }
   ],
   "source": [
    "from random import randint\n",
    "\n",
    "def display_loss_stats():\n",
    "    losses = absolute_loss(my_weights, data_X)\n",
    "    print(np.min(losses), np.max(losses), np.mean(losses))\n",
    "\n",
    "# evaluate results and compare with labels\n",
    "print(my_weights)\n",
    "evals = []\n",
    "for i in range(m):\n",
    "    evals.append(predict(my_weights, data_X[i]))\n",
    "# quickly display n random comparisons using updated weights\n",
    "for i in range(5):\n",
    "    r = randint(0, m)\n",
    "    print(\"prediction {} [{}] :\".format(i, r), predict(my_weights, data_X[r]), data_y[r], end='')\n",
    "    print(\"(absolute loss = {})\".format(np.abs(predict(my_weights, data_X[r])-data_y[r])))\n",
    "\n",
    "print(np.sum(absolute_loss(my_weights, data_X)))\n",
    "display_loss_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "adbcbca7",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Predictor():\n",
    "    def __slope_and_yintercept(self, X, y):\n",
    "        if len(X) != len(y):\n",
    "            print(len(X), \"!=\", len(y))\n",
    "            return\n",
    "        n = int((len(X)+len(y))/2) #they will be same length anyway\n",
    "        hat = lambda data : (sum([i for i in data]))/len(data)\n",
    "        xhat, yhat = hat(X), hat(y)\n",
    "        # slope = m1/m2\n",
    "        m1 = sum([((X[i]-xhat)*(y[i]-yhat)) for i in range(n)])\n",
    "        m2 = sum([((X[i]-xhat)**2) for i in range(n)])\n",
    "        m = m1/m2    \n",
    "        # y intercept\n",
    "        c = yhat-(m*xhat)  \n",
    "        return m, c\n",
    "    \n",
    "    def fit(self, X, y):\n",
    "        self._m, self._c = self.__slope_and_yintercept(X, y)\n",
    "        \n",
    "    def predict(self, x):\n",
    "        return (self._m*x)+self._c"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "678d531f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-3327.9681643574127\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fe95b2e7c10>"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAD4CAYAAAAD6PrjAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAABDDklEQVR4nO2df5QcxXXvv3dmd8XohyU0K/9AIMkmMnnmrTFYx5hHXmIsG6zFBKQktpOVWEtJCJKJhZ1YYMsnAicb23LioBwiFMUGFmli4h8CzGMVnqPgl/OI7VgY8D7wD2GMBMI22l0kEBq0v+77Y6ZHvT1d3dXdVd3Vs/U5Z8/u9vR0V3dX33vr3lu3iJlhsVgsFgsAFLJugMVisVjMwSoFi8VisTSwSsFisVgsDaxSsFgsFksDqxQsFovF0qAt6wYkpbOzk5csWZJ1MywWiyVXPPLII0PMvMC7PfdKYcmSJdi/f3/WzbBYLJZcQUQH/bZb95HFYrFYGlilYLFYLJYGVilYLBaLpYFVChaLxWJpYJWCxWKxWBpYpdDCVAYrWHLLEhRuLmDJLUtQGaxk3SSLxWI4Vim0KJXBCq65/xocPHYQDMbBYwdxzf3XWMXgIg9KMw9ttLQWVilEIE8v6OZ9m3Fi7MSUbSfGTmDzvs0Ztcgs0lCaSfuLijbmqc9azMAqBUnyZnkfOnYo0vY0SUNQhZ1Dt9JU0V+StjFvfdZiBlYpSJI3y3vR3EW+2+eX5mdqOaZloYedQ7fSVNFfkrbRpD5rRyz5wSoFSUy2vP3oW96Hme0zp2zrKHbgpZMvZWo5qhZUXmGz4YEN6L2nN/QcIqUp2h4VFf0laRtN6bN2xJIvrFKQRLcQUU1PVw92XrETi+cuBoGweO5izOmYg7HJsSn7pWE5ugX3wWO+5VZiCSo/YXPb/tswwROh5/BTmjPbZ6JveV/kdvihor8kbaMpfdakEYtOWmU0ZJWCJGEvqIkdoqerB89c/wwmt0zimeufwUh1xHc/nZajV3CLcAsq2XvpJ2yCcJ/DT2nuvGInerp6pI8XhAqlk7SNuhWfCO/zU2kImEorjYZyXyU1LZwXcfO+zTh07BAWzV2EvuV96OnqaXQIR0A5HcL9PRNYNHeR7wuq03KUEdxe5Sp7L6MIFT9h2NPVo+35BPUXWSqDlUTfd/bduHcjhqvDAIBSWynCVUTH7/kRyNcgMHWUHYeg0ZBJMkAGYhZbb3lg2bJlnHXpbJE1tHjuYjxz/TPpN0iA94UFasJSpYXspXBzQThCIFCTsItyL4OsUDdFKqJ/ZX+uXk5VzyrqcZIqItlnorvfpY2onxMIk1smM2hROET0CDMv82637iMFpBnQS+Km0u0y8UNkDS6eu7jh1nKfP8q99HOPeJnZPjMThZDUnajKDx/lOCpcIEF9vlwqp9bv0saU+I0KrFJQQFodQsVL640z6H4xo/q1Ze6lI3DX7FmDUltpirBZv2y9tNLTFQfSKVyjGhqi/Q8eO9h07SoUUVCfn90xO7V+lzZZxW90YJWCAtLqEHnM4og6OpEJ6LsF7nB1GNXxKnat2oVnrn8G2y/fLqX0dAYGVUw6K5D/qxnV0BDtT6Cma1cREA7q83EzzExL4PAji1G4LmxMQRFJfbEy5NFvKYP33nUv7cbAgQHfe6kqfqMzDpTkOfnFABxUxRREgd8iFX3TeaPek86tnY3AdpLjZBEDm07YmIJm0nDLmOS3VGXB+Vns/Y/3o295X+KYQxA640BJnlNQtpYz2kgaRxIF/id4QsmI9wPnfsB3e/fS7kjHyePIuBWwSiFHmOK3VOl6ifriq1KMOhVsnOfkKNmwzB0VcaTFcxf77rd47mL0nteLIhUB1EYOF515ETbv2xxJ+Q8cGIi0XYQpM7KnG0qUAhHNI6KvE9GPiehHRHQREc0nom8R0YH679Pr+xIR/T0RPUVEPySiC1zH6a3vf4CIelW0LYi8+CsdTPFbqrTgor74qhSjTgUb9Tm5lawMSa1l0bV3L+1G/+P9DRfSBE9g38/3TVH+q/esRufWzsB3RZUwN2lkPJ1QNXltG4B/ZebfJaIOADMBfArAPmb+HBHdCOBGADcAWAFgaf3nQgC3AbiQiOYD2AJgGQAG8AgRfZOZX1TUxinkZcKZF50TrmRRacFFnVCnYlJY3ONEiRtFeU5RZ2YDyaxlv2vvXtqNnY/sFJYIcTNcHQ58V1RNkuxb3ucbU8hjRk+eSBxoJqK5AB4D8CZ2HYyIfgLgXcz8CyJ6A4BvM/M5RPSP9b+/4t7P+WHmP6lvn7KfiLiB5rxMOFOB6iC4ynuXl2CiznYGTfBLGvyVefZBwe0gRG1Qea/SSOCYrugMNL8RwBEAdxDRo0T0JSKaBeB1zPyL+j6/BPC6+t8LATzr+v5z9W2i7U0Q0TVEtJ+I9h85ciRWo6eLv1JH6qVK10vWLrEkdZZUBT2DJvj1r+yPfa9ln/3GvRsjKwRA/K6ofKZpz6uxqFEKbQAuAHAbM58P4BXUXEUN6iMIZbmvzLyTmZcx87IFCxbEOkbe/JVx4x86hJlqQZ7Vi+8nNNfdtw6dWzub7nPYJDBdSjbJvZZ59pXBim/6qAxB74oV5vlFhVJ4DsBzzPy9+v9fR01J/KruNkL99wv1zw8DOMv1/TPr20TbtWBKJo8MSax9XSMiE156HaUkRidGMVwdbrrPQQIw6egrTPDHvdcyzz7IOChSEeuXrUe5VG76LM13JW8JIXknsVJg5l8CeJaIzqlvWg7gSQDfBOBkEPUCuK/+9zcBXF3PQnongGN1N9ODAC4lotPrmUqX1rdpIWu3RRSSWPsiYVagQq5fLt11ehyc+xxWZ0nF6Eu1kpUZDQfdg/6V/dh++XYMbRrC7lW7M3lXWqkkdV5QMqOZiN4G4EsAOgA8DWAtagrnqwAWATgI4APMPEJEBOBWAO8DcALAWmbeXz/OOtSylgCgj5nvCDu3KTOaHXQExkyaIWsKKoLdshU9nfvsPFvRd0ybWS4T8BXdg3KpjKFNQ6m1VYSqpAYbsG5G64xmZn6s7uN/KzNfxcwvMvMwMy9n5qXM/B5mHqnvy8z8EWY+m5m7HIVQ/+x2Zv61+k+oQjANXVZNkvhHT1cPes/zn/JxYuwENu7dmKhtWaHCLSZTZRU4dZ8da140+SvseaTtBpEZDYvcqNtWbNPaNllUPGc72oiGndGsEF0ZKkniH5XBCvof7xd+Plwd1vZy6BSC80vzfbdHSRRwhGYQfvc57ozlLARTmFvKdDeqioQQWy4jGrYgnkJ0FqyLO/yVcZHomJuhcpEY73UDwNp71zatN91R7MDtV94eWaCJ7lHQ4jyy+f/OPgUqKCk2N91Q0Y9atZBkUkTuI6sUFGLihLigiVEOOl4OFfdCJBBKbSXfNMq4fnAdE9NkJ4SZJJhM9bvrWg1uuitkWyU1BUxMc5UZZheooNzFo8IXLBr2i/LqR6ojjb+juK50uFBkS1fouPdxMNnvnjQzy8T30mSsUlCIif7ZvuV9IFDgPhM8oVwQqPAFxy2gFkfAqU4JlW27inuvInZjqt9dxbWZ+F6ajHUf5Yi4w2i6OVgpeFExrFbhkglKl6yOV4XHNsFdEBSnmORJZTGGuPfZ25dMTLPNS12svGLdRzknyfBelEIpQtbKDbLiVFhnQemS7mOXS2WU2kpYs2dNYGA9zbpWorb3r+zH5JZJTLK/oI3axjgWvl9fEo0msyz7YuropdWxSsFQvALXr2iZ7AsiElB+5QsAOUEgo6SSumSCFItz7F2rdqE6Xp1SmsIEARemFFXV3ooTu/ETtgxuum9R0p51pB5Pl6KVpmGVgoH4CVxRcFXmBREJqG0rtsUOwKVlxYUpFtUCTgWOkFyzZw0AYNeqXU1tVxX8jKNcRH2GwZEWBlpyyxLQzYQ1e9ZoCVBHuTZbH0kdVikYSJRFV2QtSz/hmsTFY4oVp0LAqUTWzefMNHcvfdl7Xm8qmTVBpbplRnbeleK8Kc+qjAPZawu751ZhREPVymsWhcgKVhXWb9yV3FStrpUUUTvSCCr7Bf6DRlDu++zMNHcvfdn/eD8uXnRx5OdRais1zlkulbFtxbbAYyRd0UzGaFFhHMiujhc2as3jCotZYkcKBiISrOVS2Zi0ujArLi3rLKscdJF1KhvkVuF+c9rgdi1Wx6uh30syQqwMVqSKCKoyDmTiUkGjVhusjo5VCikiKyiDsm6yXsPAIUiwpDkRKqscdJGwcdxBXrxCMmjRnjQWUOrp6kHf8j4smruoITzDzus81zDSnhgWFHswxc2ZJ+w8hZSImnNtaskBGaLME8jrdQaVD5nZPrPpOfee14uBAwON6zw+elyYPCCbi6+6pHrYeYPSfQnUiOOk/QyDrkVU6ny6l7gA7DyFzIlq1ZmwsllcZK0zHSOKtNxWQcFa78il97xe9D/eP+U6Xzr5EjqKHb7HSLqAkozrJs4oI8i63rVqF3gLa+mrYc80aLRoS1xExyqFlAhyF7QassJKtb83TbdV2LrKboU+cGCg6TrHJscwp2OO8Pgyaz8nEXhx3CpBilCX0RIlm8vPiLIlLqJjlUJKiF4oAsUWWlmm2gWdW1ZYqfb3phlUjCJsRNczUh0JnG0eptSSCLw4o4wsrG6ZZyozksjrqDsLrFJICVFhOgbHElpZVrUMO7essJJZPzqK4ks7qCgrbIIEcNK1n+MKvDgCPgurO+yZmlzdNa/YQHOKiArTxSk6lmXRN5Xr5gatH+344mWDoSYUwvMjLKib1drPeQjyBwW3F89dLAzYZ/3M84ANNBtA3LV9/cgy1U7VuR3L0y+N88TYCex8ZGckd1DcZTKdcg1tn2kD3UzKXXFhFrZj7YtqUYmWHlXRLtPdKkEjqaTlXyz+WKWQIip9sqoKqsVB5bl7unqEFUP9SksD4hc+qnvDW67BOZ8OF0RSAex1o214YMO0KN3gfqZRyLK6a96xSiFFVPpks0y1S3JuvxiB6AWWnQjmJorwDSrXoCpAHSUm4l45zs1wdbjJb37b/tumjR/deaZhi0U52JTTZFilkDKqhuxZptrFPbcoKNi9tNtXyVzz9mu0Kr4wF0NSF0TUIGiQcgyrNXRi7AQ27t2YqL2mk1X5l+lWUM8qhRzjXlMAQGORGZlOm7Sjx1FuovTCgQMDvkpm++XbtSq+MBdDUhdE1BRZ0QhM5EbzMlwdbmmBlUX5l+mY3WSzj3JOnHIFWS1zmKQsgw4qgxWs3rNa+PnuVbsT3Y841yuqvCo7ydGddZOH7KKoeK+pe2n3lPIhqq/R1Iw2FdjsoxYlzoStrCpHqlg0ReVQvqerR5jxUy6VEwuXOAF5vxFY2FwGN3Hy9/PkHnHfn77lfU3lQ1Rb8dOxoJ5VCjknTqfNqqMnXTRlwwMblA/lRavPbVuxLbGwVJUM4BfDmdU+y3dfR+HIKv7KYAXr7ls35Z6uu2+d0YrBIQ3jJsssv6ywSiHnxOm0WXV02QC16GWPOm8hSZsAJFZASdctcCskAFNGEP94xT8GKhxZxb9x70aMToxO2TY6MZpK0Dqp0k3DuJmOBfVsTCHn5CmmIEtQWWo/dMQksvQlyz6foJiBbPtFs+wBgLfokw2ia/SWGA+KEeh+Ru6Z5kUqYoInMikNrgsbU2hR4lijpleOVDlvIS5Z+pJl3SJBGWAqLFydMQbRNe7Yv0N6dKbTiveb2OiugpsmaU9cVDZSIKIigP0ADjPz+4nojQDuBlAG8AiANcw8SkQzANwF4O0AhgF8kJmfqR/jkwD+EMAEgI8y84Nh553uI4VWJMiKjFILKQlJFwoCwtcWFqEqS0sm+6hza6ewVISDjnscZTQYZPnryrAyJesoqD6YQ9znIxopqFQKHwewDMBr6krhqwD2MPPdRLQDwOPMfBsRbQDwVma+log+BGAlM3+QiN4C4CsA3gHgDAD/BuDNzMFJ2lYptCailz2tNMsoLhzvfh3FDjAzxibHAr8rIk2BVBmsYO29a6e01Q/V5w4qdOcli5RlU9KnZe9TnOej1X1ERGcCuBzAl+r/E4B3A/h6fZd+AFfV/76y/j/qny+v738lgLuZ+SQz/xzAU6gpCEuL4xdwDFo0ReVEJVGwM0lQfHRitEnIRgmIpxnc7OnqwW8u/s3Q/VS7zfyuUVTGIotMH1OyjmTvu8rnoyqmcAuATQAcFVoGcJSZx+v/PwdgYf3vhQCeBYD658fq+ze2+3xnCkR0DRHtJ6L9R44cUXQJliioym03fV2IMAUU5WWU3TftmM+3n/l26D5JhKFI6Xuv8dpl1xqT6WNK1pHsfVeprBIrBSJ6P4AXmPkRBe2Rgpl3MvMyZl62YMGCtE5rqaNSkGc1kU7VuaO8jFH2FSmkMGXs/rxzayc6t3aGKu6wMhpJhGFQX/Feo+6yJlEwJRlDZuKiamWVOKZARJ8FsAbAOIDTALwGwD0ALgPwemYeJ6KLANzEzJcR0YP1v79DRG0AfglgAYAbAYCZP1s/bmO/oPObHlNoxVIDKn3eWfpuVZxbR0whyrm8i/UEBSVFKZ+99/QKFUPSFExTArZ5RldpD+2B5vpJ3gXgz+uB5q8B+IYr0PxDZt5ORB8B0OUKNK9i5g8Q0bkA/hmnAs37ACzNc6DZ9PkAcVEpyJMKjSRKV+bcMsdXnX0Ut70yQUkCTXl2M9tn4qIzL8K+n+9r2nf9svXYfvn2RG02JWBraSaLeQo3APg4ET2FWszgy/XtXwZQrm//OE6NEJ4A8FUATwL4VwAfCVMIppOla0QnKoNwSddmSOLGCju37PH9XD06VjULmzshE7PwCugTYyfw1MhTWL9sfWMeSJGK0gohzJ1lSsDWIo+d0ayRVrWSVI+A4lr7KlwTKmYFq7gWGUTtKVIRkzyJAhWky2y7idsfZfpBq46WWwHRSKEti8ZMFxbNXeT7EreClVRqKzVe9HKpjG0rtiVaMCjOd2VnHQcJ6qBzR53V7BWAzsjCOU9S+pb3+cYMHEUQRyEA8ftj0EjYfX+dfVsprtbK2DIXGjEhrc1veJ8kndQRfO5ZsNXxqo6mhyLjmkjiYorq+tDtLvRmxIjKfhSpCAKhXCqjQMGveJL+KKs0dbjSLPqwSkEjWae1+QnEtfeubSqVHMUPn2acJEx5ySjdoPaqOL6bNOoluQXsJPu7fCZ5EpNbJjG0aQhB7uEiFRP1RxsvaE2sUtCMKispjnXvJxDHJseaSiVHEeppFYqTsfBllK6oXc7xkh7fjUgYMlhL4TIZoRwkoPtX9icyUEwYCVvUY5VCDojrAtEx2zYt61BFpdCgdhWpqOT4boImGumYqS0jlPuW96Gj2NH03UL91U/iSkw6Es7Tim/TCZt9lAPiZtlEKToWZV5AGtkkKiuF+rVXNMEraWaYuwa/H6onbQUF0cPaUi6VUR2vZpIZZLOSsseup5AyKq2guC4bP0uyvdDeZDlGGfKnFSdRNSIRtVe0NvP80vzIbfWe75nrnxEWd1PtZgsqh+FeD8CP4epwrkuMWPRglYIGVBd5kxWQfks4egXiHVfdgduvvD2RUE8jm0SlvzpKe0eqI0rcGFkHYf2ErixpLCSU5SJGlmCsUtCAaitIRkCKFBGAVGbbOm1QNTrSPSIZqY74bmewEt+/7DNLer9Ex0giXNNQXFkrTYsYqxQ0oNoKkhGQWQ/HdZTAdiuvvuV92Lxvs7KgZJDwCbtvMsI87JmpuF9Bx4grXNPKHrKZS+ZiA80ayKIyZFhgVne1Vp3XrCMoGVZRVBRwVtUWFfcr6Bii2c9eyqUyZnfMzmRx+lasIJwnbKA5RbKwgoKG42ksZKPTR6xjFORY8qJZwXFnLcuMIiqDFWEAWEUa8aFjh5pGKuVSGe2F9in7zWyfiW0rtjX6q1MmI62FjuxMZzOxSkEDWcxkDlJEcYVqFJ+3Th+xLoXT09WD/pX9ymYti5Tvhgc2TFn4Zt1964RtUrFoj7PdLXSHNg3hjqvu8O2TWbse42LnOejBuo9aCNFwXORaAsSLqER1k+jMO9ftjovixghqCwDfz7xrGIiIer9U3fM8VvO18xySk8oiO1lglUI4YZPY/F4mk8pGmyQAgtqyZs8aKeEvYv2y9ZFX1FJxz/O4OpquNk+nOIdVCjlEVQcNC6oCzS+TadZjGi+r7DlE+0WZQe7Fzi6Oho7+mcf7kASrFHKGSJDHXbsgrOSB92XKo/UoQnZJzaQCwe8Ysq6jcqk8pRy5Q1r3O28WcufWTuX3q5X6vAw2+yhniGakDleHse6+dZGDak7Q0fF9e/EGLU3JI08aTJTNvFIRbPVLMLh22bXCInkO5VJZOJkurRm+ecoEqgxW8NLJl5q2dxQ7EvVPO8u6hlUKhhLUEUcnRrFx78ZYx5UV9lmvBQGomeAlm0KqIkUUaBau2y/f3riPAJpqIjlpoXFLmUzHjJvN+zZjbHKsafucjjmJ+qedZV3DKgVDCeuIfkNnGaII+6TWY1IBpsJ6l00hFaFCIDj3kbcwdq3a5Xvvk5QyUaUY8qJwRM9UNNqSxZTRcdbYmIKhyASHeYu5zy7Ivy47Y1ZFMDFOCqlD2kHGML9+3maN60L3fUgSW8lTbMbGFHKGY9GL1tgVlX42BT8r3xHwshauiuF8kPUX5BrKwl3mjBgWzV2EQ8cONZYMdcjTrHGdow6dFn2S0XHQSC4vozDAKgWj6enqwV0r72oqT9BeaMe2FdsyapUcYYJKRuCoePmD3GUi5eJYnFnNgRC5h/Iya9zvOlbvWY3Zfz1biTA0Id7lh0ixbty7UXuZGZVYpWAQftZET1dPU3mCO666I/MXIAwZQRUmcFS9/CLrzzQfcpi1rrO9KhWOKHPulbFXsPbetcoUg2nl30X9OcvFjOJgYwoGUBmsYOPejU3BY1N9ujLEmTCnow0y8xNM8QHLxFDyMGs8qKwKYG7ef9J7EHXyYtZlROzkNUMJE55OUNYUwRUF94Q57yQuv5fNLfCcZTFHqiOxA355CZw6ZD15SpXCCROOWQtDEUnvv6jPldpKmU5MFGGVgqHIWBfeheZNF25+hAmcMOUY9ZqzErBJBGseFZkflcFKYB0oZw0H04wcFdlufs8fgJHP1SoFQwkbajsLn3jJ2spQjYxyjHLNWdRuUlUqQ0VKZBaL5rjZ8MAG3Lb/tqbtRSqiQIUpk89MEJCA2amuOrBKwVCChKF3hOBGtXDLutOGKUdA3fwEXcrUBPePaLQVVvZcV5zCHStz0qhNdKUArTNSk8XOUzAUv4wSoPYCucsjeFE59T6NldnCkLkeVfMTdKG7dk5YZowo6wcQZ7uI0kc7t3Ymfv49XT0Y2jQE3sLgLYyhTUOZ13gKwtRU17RJrBSI6CwieoiIniSiJ4hoY337fCL6FhEdqP8+vb6diOjviegpIvohEV3gOlZvff8DRNSbtG15wK8j7l61G0ObhqTLHyTFhJW3upd2B36ucn5CHGRSFXXOI5BR3GGC1fncfS299/QKCy/qMAxMry+Up8KAukjsPiKiNwB4AzP/gIjmAHgEwFUAPgxghJk/R0Q3AjidmW8gom4AfwqgG8CFALYx84VENB/AfgDLAHD9OG9n5heDzp9395EMul07JqydEORGy8In7kbWrZD16nNhcRnnPoalCouOrwIT4i6WGtrcR8z8C2b+Qf3vlwH8CMBCAFcC6K/v1o+aokB9+11c47sA5tUVy2UAvsXMI3VF8C0A70vavlZAt/Wi2nqLMwEoyMqVXYFMVxkB2ZGUTveDjGtK5IoEgtfrjnPeuCS9Rya4OlsdpTEFIloC4HwA3wPwOmb+Rf2jXwJ4Xf3vhQCedX3tufo20Xa/81xDRPuJaP+RI0fUXUCLISsoVbqo4r60QQoozI2lW1BEiRXoUuAyitstcIFapg8wtY5TVCG/aO4i5Qo3yT0ywdXZ6ihTCkQ0G8A3AFzPzFNWwOCaj0pZmhMz72TmZcy8bMGCBaoOayyyL6V7v86tnVh33zopQanSwo370gYpoDBBFvWcUYWcCX7wKOtgOGW6x/9iHLyFpwheUZv9Ci/ObJ+J7qXdRlnmdiEc/ShRCkTUjppCqDDznvrmX9XdQk7c4YX69sMAznJ9/cz6NtH2aY2sFezdb7g6jNGJ0Sn7BAlKVRZu3Je2p6tHWPmVwYHCO8o544wqTKiRpEpxi67lrpV3Yfeq3U3HHzgwYJRlboKCbnVUZB8RgC8D+BEzf9H10TcBOBlEvQDuc22/up6F9E4Ax+pupgcBXEpEp9czlS6tb5vWyFrBsr7iIOGswk2Q5KXdtmKb0CceJLydkhgy54wzkjElVVGF4g66Fr/jm2aZm6CgWx0VI4WLAawB8G4ieqz+0w3gcwDeS0QHALyn/j8ADAB4GsBTAP4JwAYAYOYRAH8J4Pv1n8/Ut01rZF9K2ZdUJJxV+eWTvLRen7gXP+Eddb3eJCOZVklVjHItUZW87nUD0lDQeVr7QAcqso/+LzMTM7+Vmd9W/xlg5mFmXs7MS5n5PY6Ar2cdfYSZz2bmLmbe7zrW7cz8a/WfO5K2rRWQtYJlh88i4awigOekCp4YO+Eb5JTBEVjetYwdvMI76nq91v0QjShKPq3MoLDFiJJgs5vsjGajiWIFB6UjOpRLZaFwTuom8K53PMETDeERx4qTFd5R1+uNM5KJE+hvFQszimWeVmaQTsFts5usUjCaKFaw1/XitbRnts8MXK0tqQWt+mWSXchetFypqN1R3Q9xA/1xBJWpSkXW3ZRW/EGn4DYthpIFVikYTFQr2J2OuGvVrkh+16QBPNUvU5jwdoSwXwXZsHZH8aknCfRHEVStsL5vFMNCxwpnKgS3dS9apWA0STpo1MBo0gCejpcp6BpE2VZFKioNPCYN9IvSYr0CMcr6vmv2rAHdTFoURBJhLWtYJB1V6RTcNrvJKgWjSbuDJsmwSbutIiE8yZNKM1FkBZDsfiKBKKpZ5Le+r1OnSnUQNKmwFhkWANC5tRN0M4FuJlx9z9WJRlU6+5op6cdZYpWCweSlg6rIOoqKjBBW4XaRFUCy+4lGBM59i4rKIKgKX73XsACAtfeunbKGwiT7F1mUdf/ofi9aKf04DnaRHUsiKoMVrLtv3ZTZ0x3FDtx+5e1aX6awapsqK5bKVuWU2S9oMSG/daxF6/t6v6eimq2OarlRFrM3YaGd6YRdec2ihc6tnb5Cq1wqY2jTUOTjRSmLHLRv1qugiZBZ1J7BjTLXQPP6vl5UXZOO8uUyK+oBrb3CmamIlEJbFo2xtA4iKzbMuvVjwwMbsGP/jiafOQBfYeGUZvDD1NTCsPUMHIXgFfLOust+owkZX7pXgXYv7cbAgYEpCjWobWHPQsSiuYuEiqZIRUzypF0TwTBsTMGiHRnffmWwMkUhOJwYO4Hee3ojxwVMTS0MK+UBNCuuJKnGgH8A+bb9tzUFlAFELjMSRt/yPrQX2pu2dxQ70L+yf9r67U3Guo8siQhzH8n4/h0rWAZZN0MeFmFPy8Ul69d3n1dlfKEyWMHGvRsb/aRcKmPbim3GPIfpiraV1/JIXiYD5YFtK7Y1WYLthfbG7OmgjBZvaQwZZK1VUzO33H1v6IR/zCVsveqoyLrM3PupHGn1dPVgaNMQeAuDt3Bj/XGLmUy7mILXgozrK7XUcO6ZKOAb5NuPujSk+7uybTPpmXr73itjr/juN3BgQOl5g/z63v0c+pb3+WaVTadJXNOVaTdSsAWv1BOU1x1kccYN+mYdF4iLrBL0E+CqZxp78QtYe13LMq5mOwrPP9NOKZialdKqiCZ1dS/tFhazA2r+7fXL1rdUyQHZPuaeyFYZrKBzaydW71mtdKbx+mXrA11rfsUYxybHAo0nW3a6NZh2gWZT89dbGW+gcVb7LIxNjjUtFwo0B4OjzFswCb92Rwmo8xb2DZa70dln4wSa7buVL2yguY4teJUN1fFq4+9Xxl7xVQgFKjRZrHksOSCymLuXdoe6cQA0UkLD3E1RAvRRiRNoFo2EDh47aEcLOWLaKQVTs1J0krWfV9aXLqqJE0bW1+dFFLcaODCAnVfsRLlUFn7XbaCEuZsIpO1a4xhPQQrDupHyw7RTCoD51qdKIWeCnzdKvCZqwN+E6/MSFLdy0jN3r9rdGBGIigiGBdQZrC1BIo7xFBTQVpnMYZoR0GpMu5iC6aiedCXj59Xtt49SFC3q5CgT/diq2hQWUwDUFcNTRWWwgtV7Vvt+5m5r3D6Xh0mJecHGFHKC6pTZsGyrNCxtmZRIh6jppiZmk6mKW8mUxDAtPbenq0fYXqetSfpc3PdD9ei7lUcqViloIm7HCQrWxTleWMAwjXkbopRIFYJTdH0FKmT20qqMWzmuzt2rducmQSJMKSbpc3GMAJWGj4nuStVYpaCBJB1HJOQIFOt4YS9oWpa2N46z/fLtoYJTRrGKRiETPJHpS6s6bqU7QUKl9RvW1iR9Lk5WlErDZzpMfrUxBQ0k8Sn7+Uy95ZLDjidTJtn0dQei+I7d11ugAiZ4oul4WV+Pyfjd6wJm4KMX/A3+7orrlJ9P9fsRFlNQWdxPx0JEDmnPybGL7KRI0o7j7RyiIK3f8aK+NEH7A+KaRrqJKjjuffQwvvDgT/Cfr74HID0vrXOO549Wcca8Ej5x2Tm46vyFkffJEnff6iydgXmjV+PpsS9honCkad82fi2+duUPprQ/7Prcn88ttYMIOHpibMq+voYPz8D8setwzpzu0Hv2sftvxT/84C8xxkdQ5E6cPn41Zk1cgiIRJpix0NOupIaP+5p+UVqHk3jB91i3/NbDuPn+J/DiidpM8Hmldtz02+dKPf8sAuhWKaSIaus7yvHinNvPQgGaV/xS0Un9hNLY8YubBIyMYnVe1sNHqyAADOC5GWt9BVxxcgEuLN2NT1x2DgBEFtz3PnoYn9wziOrYqVGIc05HCAGYss/x4kM42nYXJmgIRe7EvPGrcdaMS7HlinMBAJ/42mMY8+ip02e2Y8sVzYJEJIw/fe8gvvK9ZzHheo8d4ej8LhAwybX2jLTfCqaTp66BZ4BxsnYxXpiw+NX7Ma/UjtHxCZzwNhZAqb2I33n7Qjz04yM4fLTqc5BmZrYXMIJ/x1Chf8q9ATDlfr2p/Y+w9fKPYP/BEfzz9w5hMoKoai8QvvB75zWU0B/e98c4OXGqfTOKJaw/byv+64m3TLmnYefyu4fOsf7Xd9+EsQn/L84TKEiHKO+tKsPDKoUUUa31owzvxcsfEpZU7w/tRE6H+171Q77CNcxSD+qofi+nYyHOnrgEpfYiPruqCwDwe/ddgHFqtsiKkwuwePTOKULQjUjwOedoLxBAEL687QVgnAFmoEDAjLYCXh2bRKEuYJ1zeIX9nIlLptz1sHaIcB97Bi3A+876GH709Hk4Od4skAsAoox9RAoTXACo+UjFyQU48+QdEc4QH9H9Ko9dh1kB9yuIeaV2PLblUnz63kHs+H7/lGc2f6IXcycvwZhL+sveT+/zfy1/GK9re09jhCCD09dljKA9v/3slNHXK6PjU/qv91iyWKXgQscQ33vMd5z7JO752d80uV7c+82b2Q5m4FjV33pw87H7b8Wtj3wG4zgliE7ndzesIQD49L2D+PyjyzEusJSdF5wA9LxzEf7qqq4p+3z63kFUvnsIDODgaVcEumHuffQwbvrmEzhaDX8RZnUUMcmMA4VeoRXvFj5FIvyq7R/wSnFgigUrI1QBf6Ed9h1ZZIV90IhFJGjjKhJZRM8UDBBmaDuvDGH3K+4znVdql+qjWbBwXgkP3/huAMBrt56JI9XDTfu8pv0NOKN6+5QRatixZLFKoY6fGyBM08r4Ub3H9KOjSBgVWKgOfu6Dex89jI/9y2O+9r/bGtr93UPKBEvQS7p0sh9VH1dCGGKhVHNTOPhdAxiYNdGNzrENkc+rEllhHySAi7wAEzQE4tkgAiZxHEXuBNOrmKSXQ4+to+3zxq/WpkhlCOob5bGPa1WWWUEAfv65ywEAb/6rT+Gp8S82XWPn2J9i5sS7Ih1L+vxWKdS4+HP/jsNHq3jb8z/B5/dua2wnEF4/9zS85rQ2gBkvvPQqjnksDGKe4notENA5ZwZePH4S4xPNQpLAvv28tt3/vlP9PM3IH4vpJBgnQDyJM477nsZiseScu996KW5c8VEAakcKxq28RkTvA7ANQBHAl5j5cyqP/3w9GHbD/7kT5wx58qJdqyO+tv4TyjAwR1XjLBaLRZKVTzyEG1d8FKX2YiPRQQVGKQUiKgL4BwDvBfAcgO8T0TeZ+UlV5zhjXgmHj1Zx9Qc+gzeONPvwXjvnNAwdH8W4z3cZ5JulwSAwNX/AgO/2sO/AdzuBRef2TR0BmIDZo9Wmz5nIt21Om3zHKVTwPf8k/Lcfb/sOjrZ9HRM0jCKXMXf89zBr4mIcLz6Mo+1fw3hj+wcwa/I3asei2rU8P+M6TNDwqXNQ7VqKk2UsHN3uOrf/PXSO439Pprb3uRl/jAkaapzDoTjZiTNPfhmTVJjyPKK4544XH8KL7TsxiZf9s3t8G4hYcZQghAFmD27/fRbuGlHcIE58JowCgD945yI89OMj0tlHToopIJe9dv5n/rdv8LlIhL/9wHmhLmgns+sbjxyesr29QJh9WhuOnhhrSr9VgVFKAcA7ADzFzE8DABHdDeBKAMqUwicuO6d28wH8dMGSps8PAODTVJ0te36VwTlrQuVOl1AZxi/4Dswfm4XZE+/HDLwfM+qfnGwDTnq+z8UP46ivUPowXqZZStv6atuwr8CepGFMFopN2x2hKON/nz1xCY623QUUmuMEIojnoMCnKfXtT9BQ+E6u/Y623TU1noOaS/Jo211alcLsiUt8jz9v/GpfJeWksQJAx+yHcbTjLgxVn0eRF2Du2JrGsdoLhI62Al4ZrQnWoPkDV52/sCkBww8ZIXxUkI00ydz0fed/P2WzbPH8VOe+mKYUFgJ41vX/cwAu9O5ERNcAuAYAFi2KVhDMuZl/9tXHfdMaz5hXwi+PvSpMefSj6EpXtCQXKlEEb1KK3IkJ8rFCuTOwfbJtkRXIQF3xjV+j9DqPFx/CqRkVwTjXLGqzs/3is+fjP382MuWIBOB/nD0f33l6xNfKbi+gMSfj9JnteMsb5uC7T7+ICWacKH4bJ2dV8PLYL9GBBZgzugZvntONJeUSHv7ZSFN/aMMCzBur7fOJy87BK23fxjX3b8OJai1le5xewIsdt4JG0dgniwmEjlfCb7sfV52/UKio0my/aUpBCmbeCWAnUAs0R/2+c4P9hmvOENKb1ywSSu0FwgffcVbTEE8W52V64vmXjUqdkxMj/oQJFRlEgve0IuHVkAyuKMhYoUkQKZ0muBDqnim1FzA+MTllwtusjiJOjE6g3cdSBoCR9lt95yB4cV+zqM1tWIBbPvg2XHX+QmFGnjdVWTQZz6E2B2d7Yw7OSbyA4qzbsOHy89HT1dOYnDd74hLMnXw3fv/Cs5os+SW3NNcjmsRJzFrwdTx8/d+EXrsuGl4JHxljMkZlHxHRRQBuYubL6v9/EgCY+bOi7ySZvCbq2JXBCj58zx9hnF9t7NtGp+G687+Abz/6Zt9p7O7ZtX4jh4vPno/KH18UmN7qPoYfXp/r6/BhLCi8B0erY41ZqyIIQJvLWgtiVkcRfSu7sP/gSGO2bJEI73zT6XhmuNpo+yW/vqAxk9V9fpEPuMBzQBFdI95Zw16BVGovoDo22VBgHUVCW4F8Z9+6mdlem5Q2CX9f9tzJdzeu1/1MF84rYUm51LBy3bTXy0u6T/1K8SEMe9Nrfa+zNkkpTpkE0eRGcIdviquXdrwWrxmtuVvmldpxyQU/xc7BG1IpuaBi9r/OekRJMbnsSS5SUomoDcBPASwHcBjA9wH8ATM/IfqOjhnNOovEmbC4iO6O6tfWjmIHmBljk6dGQzqLrqWB7LPUXbAvyiJGbsJqYqVR90qFQDe1qKPp5EIpAAARdQO4BbWU1NuZObBgvA6loMvySCLY89bxvULl+OhxDFeHm/Yztf1hxH2WOgqfiUubBLN71W5pgySpghAdQ0W/tquxxSM3K68x8wAzv5mZzw5TCLoQ1WZncKJa81ksLpLVClHe9QRGqiO++4nab/rqVnGfpY51EUT9tVwqgwS5sIvnLo40Qk2yqEzQMVSsUqd7rYnphnFKwQSClo9MsmhLmouLmLZCVJT2R2l7VsojybPs6epB3/I+LJq7CIeOHcLmfZsTtVskWLet2IZrl13bpBiiCF0Vi8oEHUOVQFe9qNF0xioFH8LWxo270lKcVaMcolpUpq0QFaX9sm3PUvEleZaq2x0kWLdfvh27Vu2KLXRVrMwXdgzVAt2kUaZJbZHFKgUBTkcVDb/jLFeZZKgc1aIybUH7KO2XaXtlsILee3ozU3xJnqUOhS0SrEnjAUmUn8pjyGLSCNmktkTBKoUQVHbopEPlKBZVmi+iLLLtD2u787L5ZfEA6Si+JM9Sp8J2W6adWzux7r51iYSSCp+/imPIEkXh6rbiTRuty2KVQgiqO3Ravs80X0TVhLXd72VzM780X2v7HOI+S10K22uZDleHMToxOmWfqEIpqSFTGaxg496NU55XuVTWFgiWVbhpWPGmjdZlsUohhLxmNuS13UB4201/qcLQpbDDlKVD1PsXV/lVBitYe+/aplTkl0dfbnye1FL3HkNkEHgVbhpWvImjdRmMm6cQFROX47TIEdffHTZZyz2fJK1JWFHR0S7Z+QppzQ0Jek7lUhnV8WqiuQV+8xPaC+0goikjJL/jpjEL2vT5E7mZp2BJh6yzIpIM34NShoHm2IOJgT4dbkRZC/T46PFU7kHQiGS4Oqwl1XVscgxzOuaEjpDTsOLzOlq3SkETWQvdIEwQlkmG787LVi6Vmz4Liz2kGehLuw/4Kcv2Qjtmd8yesm24OpzKvI84AlZFqutIdSRU4aYVc8vj/AmrFDRggtANImthCYhf6IPHDkoJo56uHgxtGsLuVbubLDEg2HWRRkwiiz7gZ5necdUdvsozjXkffcv70F5ob9reUezwbROQXqprXq34NLAxBQ2YXqfIhKqSYXGBOP7lzfs24+CxgyBQoG89jedgSh+oDFawes9q38+8z1tHm53sIyfYXC6VsW1FbW30pP520332pmNjCilieiqaCVkRYXGBKCMXt4ULIFAhEAjdS7ujNTYGQX0gLbeSc19EeJ+3jn7rjOh4C4O3MIY2DaGnq0eJpW6tfT3YkYIGTLESRZhiYbmtexEECs3OiVo6emb7TPSe14uBAwPaspJEbVKRdZO0DaJzivYvUhH9K/utsG0x7EghRUyfOKbSwkpi9TpBOFGNKQBSvu2oluyJsRPYsX+HVn+/qA845/e2R0c8J+i++D1v0ehtgid874/JyRSW+FiloIEwoZum+0B0HhVZEaoCk2GuJCBYcMZxe3ldTKoFs6gPRC0hngTRfRGVzXbaXKRi02fe+2N6MoUlPtZ9lDJpuW7SOI9KN5l7MpcoJiAKhPtdqxNsLlJRWCNJ9viy7ZZxQ6XpWhT1gTDXmUwigukuUks41n1kCLrTQZ3Rweo9q7W7KVSmfLpHLiJ3ksjy9bPKd63aBd7C6F/Z3zQKEVW+jTriiGMtp+la9Lsvvef1ov/x/sA2yyQimJ5MYYmPVQopo/Nl2vDABqzZsyYw6Krqpd3wwAbhZ0mzmOIITpE7zE8wXrvs2qbjdxQ7cHz0eCSXXhwFn3bGjPe+DBwYCG2zzP03IYPNogerFCKSNB6gs0Lmjv07QmvfqHppdz6yU/hZUqtXt+C8eNHFU45fLpXBXKsqGsU/HlfBZznLVabNMvff9GQKS3xsTCECKvz0unz9nVs7m6pRelEZU6Cb/V0wQM2vbFLhubD7Hdc/nke/uq44kEnFBi1y2JiCAlTEA3RYwZXBSqhCUG1t+2WoOJiUiSLzzKJa/M5o0Zk97SYtaznuiFWlhZ/Huj6WcKxSiICqeECS+vR+giBIKREIu1ftTpR26nfOa94unikLyCnLNFJzZZ5ZFJee3+xpRzGkNaM2STqonQVsCaMt6wbkiUVzF/kOvdMIrnndII4gAIKV0rXLro39wgedc/vl2wHUYgtxlsUMOrZKASV6ZgzGkluWoG95H/qW9/mmtPqVw/AbeTA4VZdR0OhH5t45ZSYsFj/sSCECWQbXggSBSCmVS+WG8FZ9TqCmGMb/YjxyCqnMsVURNDHOrYh6z+ud4gpiMPof72+yvuO6mmRHQzL723RQi06sUohAlkPvIEEgUlZONUod53QTR1mmJdjcz8wPRxENHBiQmuUc19Uk4+aR3V/UhgIVMi05YctetAZWKUQkq+BakDDSpaxk17sFgFJbqfG3zMLsqlNzZUp6iCatHTp2SIsCjDoakt0/qEZRViUnbNmL1sEqhZwQJoxUK6vKYAUvnXypaXtHsWOKAHSEgTv7qTpeDT2+SldcUgt70dxF0tZ3FAUcpmi8ikx2hri3DTK1inRjwsJNFjXYeQo5Is288KDSz0ObhkL3kwm8eq+ne2l3rHLWsm0ImrMANC/64iXqPI+gdomC236TD8PuZZxFk1T3JRMWbrJEw85TaAF0ua78XC9B69+6SbKspvt6+pb3hdbkESHr+gmy8nVY30GjIVEWU5x5D1FdcTpcPbbsRetglYLh6A7eiQSEbDwh6KWPImySuB+iCKQgxer+bJL9rdsogfAgJSQ6jlPZFTg17wFAYB+I6orT4eqxZS9ah0RKgYi+QEQ/JqIfEtE9RDTP9dkniegpIvoJEV3m2v6++raniOhG1/Y3EtH36tv/hYg6krQtKSZkUqi06IImvvkJCABSL7mqZTWTZCPpEEiqLF+REgo6zgRPTGl/WB+Immiga9lNOymuNUgUUyCiSwH8OzOPE9HnAYCZbyCitwD4CoB3ADgDwL8BeHP9az8F8F4AzwH4PoDfZ+YnieirAPYw891EtAPA48x8W1gbdMQUTFmuMq6/3s9X3/94v+/1rNmzRugL3rVql5TfOWxZTRm/ctKaPKp95DJ9IMk5/Y7vxUmjVV1fKY81myzqEcUUlAWaiWglgN9l5h4i+iQAMPNn6589COCm+q43MfNl9e2frG/7HIAjAF5fVzAXufcLQodSMOWliRtAjBLABNQJnaRB57iKWFcAPui4cRew8Tt+kDIFmleJcz6LG8A1xeixZEsageZ1APbW/14I4FnXZ8/Vt4m2lwEcZeZxz3ZfiOgaItpPRPuPHDmiqPmnMGXGaBwXhiiA6UfQxLc4rpckx4rrfghzsalYQ9ov/iByu3nXfl69ZzU6t3b6njdsjeqgVNkkAVzr6rEEEVr7iIj+DcDrfT7azMz31ffZDGAcQCqOd2beCWAnUBspqD5+ljWO3PilLcadKeyHM/ENgBJLO86xklr5IuHce08vHj708BS3mcr6SkGBYi/D1eHA84Y957A+EOce2vpHFhGhIwVmfg8z/3efH0chfBjA+wH08Clf1GEAZ7kOc2Z9m2j7MIB5RNTm2Z4JaWVShFmxcSw6keIKSnX0poZu3rc5doA9StqsikC6SDhP8AR27N+hbUJVVAMh6LxRUmW9fcDOJLaoJmmg+X0Avgjgt5j5iGv7uQD+GacCzfsALAVAqAWal6Mm9L8P4A+Y+Qki+hqAb7gCzT9k5tBqbromr+meKKbLr5vE1522r1lF7CZoJnAQSRcCihK7cX+ueiKXCfEvu9hOPtESaCaipwDMQM3SB4DvMvO19c82oxZnGAdwPTPvrW/vBnALgCKA25m5r779TQDuBjAfwKMAVjPzybA25HVGs86XOe5LGjYDV/WLr2IWrEwWj4ikCk8my8uNDkGd9UxiG7TOL9qzj7LCFKUQVRBn/TL7IWoTUHvRVb/4qhRjZbCC3nt6hes6BBF0rjjKtTJYwca9G5tWwtMlKHWPFMLugQkjFUs8bJkLjcTx65pYFkB07iIVtfjmg2I3UbKGerp60L+yP3ASnYigdRDi+Op7unowtGkIu1ftTiW7R2f8S+YemJKppwsTJrGmjVUKCohTNsDEsgCiNsVZWU0GURAVaJ7Fu2bPGmx4YIPUsaIgUoRJS0E4Afddq3YBANbsWaNFqOhML5W5ByYaN6qYrkF8qxQUEMdakskqUWWhyB5L1KY4K6vJ4petJJprsWP/DqkCe6L2Rik2p8ICTkuo6CqUKHMPTDRuVDFdy4FbpaCAuCthiV5m1TWPohzLr01pv/hBcwA279scquRE7b122bXSFrUKCzjvQkXmHmQ9EU6ne6fVXWMibKBZATIZMFECjSqDdyqDuUEBR5VpiWFppjJB76TtUVF2I0ktKBMwPbNId/taPYhus4804xZCBSr4+uFlO5PKzKQ0spxUv5yVwYqwUF+RirHubdxMojjfkSl0lxehYvIchDQyr0xWikmxSiFF6Gb/tYABgLeE328TRwppn2PDAxuwY/+OwMlgbqIWCRS93EmFoMwop1WEStakZfCYqhSTYlNSU8Rv1a6g7V5MKVIniw7f6/bLt2PXql0ol8pS+0ctEujn29dZdgNA6v72VieNzKckQfy8prNapaABUQqn7OQqlcG7JMfyduoND2xo6uSVwQqI/EdGotXborR9dsfs0P3iFgn0blcRGBYJJGfUZBWCOkzOfMpzOqt1H2mgFQJUMr7xjmIHmBljk2O+n5dLZQxtGkrUhtV7Vgs/J5DUkF72eegqu9FqLiOTXComtcVNHmSAdR+liMkWjCx+VrOX0YlRoUIAgJHqSOzzO8JVxOK5i6WH9LLPQ4U7QuXIzESrUuf6FXHQNUcjKXlOZ7VKQQNZ526rQEXnTeLbDVJKURWs7PPIUpnnxd0Q5GLLyzWkQZ5nelv3kUZMHdrKELcktZtyqYxtK7bFuuag4ny7V+1WXsbceU5OHGSkOpLq/IY8uBuAYBebaHEq064hDfLgRrTuo5TJu9XkZzVHxVlxLM41BwVsdaxr4Tyn4eowquNV7Fq1K5Y7YuPejaHBaj8XS17cDUEWcF6uIQ3y7C2wSkETeSlxIPIB+3Xq9cvWRy44F/ea03LlqHxOlcFKU8lsB0cwiowFUaZWHHeDTr9+0HPJs8tEB6bGO8IIXaPZEo2wEgcmWU3eIa53DWPROr5RXUtxrlm03rNzflUuOZXWbZAicQSjSAmV2kq+5TuiKsGwZ5qUsHW4o64pbjEPO1JQiNsKFCFrNaWRxRHXShZZi6KJZnEtRa+lBTSX1E7qkguybqM+gyBF0r20O1CZjlRHlLgb0hihiizgPLtMLKewgWaFqCpxkFaQKklevl8QHfC3FE0uUBa0prV3ac2waxG1b1b7LDA4cT0kmcQFE1f0s5iJDTSngKoSB2nFI5L4gP2sRd2Woo5ApqjNAwcGlC2cdFrbaaEVdMNcLLKJC9avb0mKVQoKUVXiIK0sDh3BXJ3BNV0Cz6/NKhdOCprEJ6s4ZQ2FVpg4ackWqxQUouqFTMvay5sPOE2BF/cZ+CkYFcaCrJLK2zO1mIdVCgpR9UKmKfzylDaXpsAzrVJtFCWVp2dqMQ8baDaUPM+GbhVUPoMsV4KzWPywi+xYLDnHGgoWlVilYLFYLJYGNiXVYrFYLKFYpWCxWCyWBlYpWCwWi6WBVQoWi8ViaWCVgsVisVga5D77iIiOAEi2RFg+6QQwlHUjMsBe9/Rjul677utezMwLvBtzrxSmK0S03y+drNWx1z39mK7XntV1W/eRxWKxWBpYpWCxWCyWBlYp5JedWTcgI+x1Tz+m67Vnct02pmCxWCyWBnakYLFYLJYGVilYLBaLpYFVCgZCRLcT0QtE9P98PvszImIi6qz/P5eI7ieix4noCSJam36L1eF37UR0ExEdJqLH6j/drs8+SURPEdFPiOiybFqdnCjXTUTvJaJHiGiw/vvd2bU8GVGfd/3zRUR0nIj+PP0WqyNGX38rEX2n/p4PEtFpOtpllYKZ3Angfd6NRHQWgEsBuNdg/AiAJ5n5PADvAvC3RNSRQht1cSd8rh3A3zHz2+o/AwBARG8B8CEA59a/s52Iiqm1VC13QvK6UZvQdAUzdwHoBbArpTbq4E7IX7fDFwHs1d4y/dwJ+b7eBmA3gGuZ+VzU3vUxHY2ySsFAmPk/APit9v53ADYBcGcHMIA5REQAZte/N669kZoIuHY/rgRwNzOfZOafA3gKwDu0NU4jUa6bmR9l5ufr/z4BoEREM7Q1TiMRnzeI6CoAP0ftunNNxGu/FMAPmfnx+neHmXlCR7usUsgJRHQlgMNOp3BxK4D/BuB5AIMANjLzZNrtS4HriOiH9SH36fVtCwE869rnufq2VsLvut38DoAfMPPJtBummabrJqLZAG4AcHO2TdOO3zN/MwAmogeJ6AdEtEnXya1SyAFENBPApwD8hc/HlwF4DMAZAN4G4FYiek1qjUuH2wCcjdr1/QLA32bamvQIvG4iOhfA5wH8Seot04voum9CzbVyPJtmpYLo2tsA/AaAnvrvlUS0XEcDrFLIB2cDeCOAx4noGQBnAvgBEb0ewFoAe7jGU6gNrX89s5ZqgJl/xcwT9RHQP+GUi+gwgLNcu55Z39YSBFw3iOhMAPcAuJqZf5ZVG3UQcN0XAthafweuB/ApIroum1bqIeDanwPwH8w8xMwnAAwAuEBHG6xSyAHMPMjMr2XmJcy8BLUOcgEz/xK1oPNyACCi1wE4B8DTmTVWA0T0Bte/KwE42RrfBPAhIppBRG8EsBTAf6XdPl2IrpuI5gF4AMCNzPxwBk3Tiui6mfl/ut6BWwD8NTPfmn4L9RHQ1x8E0EVEM+tB598C8KSONrTpOKglGUT0FdSyCzqJ6DkAW5j5y4Ld/xLAnUQ0CIAA3MDMuS0z7HftAN5FRG9DLaj+DOruEmZ+goi+itrLMQ7gI7qCb7qJct0ArgPwawD+gogcl+KlzPxCmm1WQcTrbiki9vUXieiLAL5f/2yAmR/Q0i5b5sJisVgsDtZ9ZLFYLJYGVilYLBaLpYFVChaLxWJpYJWCxWKxWBpYpWCxWCyWBlYpWCwWi6WBVQoWi8ViafD/Ab8FMzTrpEoNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xx = []\n",
    "yy = []\n",
    "for i in range(m):\n",
    "    xx.append(my_weights[0] + (my_weights[1:].dot(data_X[i])))\n",
    "    yy.append(predict(my_weights, data_X[i]))\n",
    "    \n",
    "p = Predictor()\n",
    "p.fit(xx, data_y)\n",
    "\n",
    "print(p._c)\n",
    "    \n",
    "plt.scatter(xx, data_y)\n",
    "plt.plot(xx, [predict(my_weights, data_X[i]) for i in range(m)], c='red')\n",
    "plt.scatter(xx, [p.predict(i) for i in range(m)], c='green')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "adb3293b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}