Health Insurance Predictor.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 656,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import LabelEncoder,OneHotEncoder\n",
    "from sklearn.linear_model import LinearRegression\n",
    "import math\n",
    "import matplotlib.pyplot\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 657,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainingPercent = 90\n",
    "testingPercent = 10\n",
    "epochs = 10000\n",
    "learningRate = 0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 658,
   "metadata": {},
   "outputs": [],
   "source": [
    "def readData(fileName):\n",
    "    df = pd.read_csv(fileName)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 659,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   age     sex     bmi  children smoker     region      charges\n",
      "0   19  female  27.900         0    yes  southwest  16884.92400\n",
      "1   18    male  33.770         1     no  southeast   1725.55230\n",
      "2   28    male  33.000         3     no  southeast   4449.46200\n",
      "3   33    male  22.705         0     no  northwest  21984.47061\n",
      "4   32    male  28.880         0     no  northwest   3866.85520\n"
     ]
    }
   ],
   "source": [
    "df = readData(\"insurance.csv\")\n",
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 660,
   "metadata": {},
   "outputs": [],
   "source": [
    "def encodeData(df):\n",
    "    encode = LabelEncoder()\n",
    "    df = df.iloc[:,:].values\n",
    "    df[:,1] = encode.fit_transform(df[:,1])\n",
    "    df[:,4] = encode.fit_transform(df[:,4])\n",
    "    df[:,5] = encode.fit_transform(df[:,5])\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 661,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = encodeData(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 662,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[19 0 27.9 0 1 3 16884.924]\n",
      " [18 1 33.77 1 0 2 1725.5523]\n",
      " [28 1 33.0 3 0 2 4449.462]\n",
      " [33 1 22.705 0 0 1 21984.47061]\n",
      " [32 1 28.88 0 0 1 3866.8552]]\n"
     ]
    }
   ],
   "source": [
    "print(df[0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 663,
   "metadata": {},
   "outputs": [],
   "source": [
    "encode = OneHotEncoder(categorical_features=[1])\n",
    "df = encode.fit_transform(df).toarray()\n",
    "encode = OneHotEncoder(categorical_features=[5])\n",
    "df = encode.fit_transform(df).toarray()\n",
    "encode = OneHotEncoder(categorical_features=[7])\n",
    "df = encode.fit_transform(df).toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 664,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.0000000e+00 0.0000000e+00 0.0000000e+00 1.0000000e+00 0.0000000e+00\n",
      " 1.0000000e+00 1.0000000e+00 0.0000000e+00 1.9000000e+01 2.7900000e+01\n",
      " 0.0000000e+00 1.6884924e+04]\n"
     ]
    }
   ],
   "source": [
    "print(df[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 665,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 666,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    0    1    2    3    4    5    6    7     8       9    10           11\n",
      "0  0.0  0.0  0.0  1.0  0.0  1.0  1.0  0.0  19.0  27.900  0.0  16884.92400\n",
      "1  0.0  0.0  1.0  0.0  1.0  0.0  0.0  1.0  18.0  33.770  1.0   1725.55230\n",
      "2  0.0  0.0  1.0  0.0  1.0  0.0  0.0  1.0  28.0  33.000  3.0   4449.46200\n",
      "3  0.0  1.0  0.0  0.0  1.0  0.0  0.0  1.0  33.0  22.705  0.0  21984.47061\n",
      "4  0.0  1.0  0.0  0.0  1.0  0.0  0.0  1.0  32.0  28.880  0.0   3866.85520\n"
     ]
    }
   ],
   "source": [
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 667,
   "metadata": {},
   "outputs": [],
   "source": [
    "def normalizeData(df):\n",
    "    df[8] = df[8]/df[8].max()\n",
    "    df[9] = df[9]/df[9].max()\n",
    "    df[11] = df[11]/df[11].max()\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 668,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = normalizeData(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 669,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    0    1    2    3    4    5    6    7         8         9    10        11\n",
      "0  0.0  0.0  0.0  1.0  0.0  1.0  1.0  0.0  0.296875  0.525127  0.0  0.264777\n",
      "1  0.0  0.0  1.0  0.0  1.0  0.0  0.0  1.0  0.281250  0.635611  1.0  0.027059\n",
      "2  0.0  0.0  1.0  0.0  1.0  0.0  0.0  1.0  0.437500  0.621118  3.0  0.069773\n",
      "3  0.0  1.0  0.0  0.0  1.0  0.0  0.0  1.0  0.515625  0.427348  0.0  0.344744\n",
      "4  0.0  1.0  0.0  0.0  1.0  0.0  0.0  1.0  0.500000  0.543572  0.0  0.060637\n"
     ]
    }
   ],
   "source": [
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 670,
   "metadata": {},
   "outputs": [],
   "source": [
    "array = df.iloc[:,:].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 671,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.         0.         0.         1.         0.         1.\n",
      " 1.         0.         0.296875   0.52512705 0.         0.26477671]\n"
     ]
    }
   ],
   "source": [
    "print(array[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 672,
   "metadata": {},
   "outputs": [],
   "source": [
    "def shuffleData(df):\n",
    "    shuffled = df.sample(frac=1).reset_index(drop=True)\n",
    "    return shuffled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 673,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.shuffle(array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 674,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.         1.         0.         0.         1.         0.\n",
      " 1.         0.         0.453125   0.38085827 2.         0.07693863]\n"
     ]
    }
   ],
   "source": [
    "print(array[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 675,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 676,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    0    1    2    3    4    5    6    7         8         9    10        11\n",
      "0  0.0  1.0  0.0  0.0  1.0  0.0  1.0  0.0  0.453125  0.380858  2.0  0.076939\n",
      "1  0.0  0.0  1.0  0.0  1.0  0.0  0.0  1.0  0.406250  0.548654  1.0  0.045521\n",
      "2  0.0  0.0  1.0  0.0  0.0  1.0  0.0  1.0  0.421875  0.585921  1.0  0.545809\n",
      "3  1.0  0.0  0.0  0.0  0.0  1.0  1.0  0.0  0.578125  0.579334  0.0  0.584443\n",
      "4  1.0  0.0  0.0  0.0  1.0  0.0  0.0  1.0  0.453125  0.557877  1.0  0.317981\n"
     ]
    }
   ],
   "source": [
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 677,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getTrainingAndTesting(df,trainingPercent):\n",
    "    percentData = int(math.ceil((len(df))*(trainingPercent*0.01)))\n",
    "    trainingFeatures = df.iloc[0:percentData,0:11].values\n",
    "    trainingTargets = df.iloc[0:percentData,11:12].values\n",
    "    testingFeatures = df.iloc[percentData:df.shape[0],0:11].values\n",
    "    testingTargets = df.iloc[percentData:df.shape[0],11:12].values\n",
    "    return trainingFeatures,trainingTargets,testingFeatures,testingTargets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 678,
   "metadata": {},
   "outputs": [],
   "source": [
    "trF,trT,teF,teT = getTrainingAndTesting(df,trainingPercent)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 679,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.         1.         0.         0.         1.         0.\n",
      " 1.         0.         0.453125   0.38085827 2.        ]\n"
     ]
    }
   ],
   "source": [
    "print(trF[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 680,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getMatrices(x):\n",
    "    ones = np.ones([x.shape[0],1])\n",
    "    x = np.concatenate((ones,x),axis=1)\n",
    "    weights = np.zeros([1,x.shape[1]])\n",
    "    return x,weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 681,
   "metadata": {},
   "outputs": [],
   "source": [
    "x,w = getMatrices(trF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 682,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]]\n",
      "[[1.         0.         1.         ... 0.453125   0.38085827 2.        ]\n",
      " [1.         0.         0.         ... 0.40625    0.54865424 1.        ]\n",
      " [1.         0.         0.         ... 0.421875   0.58592133 1.        ]\n",
      " ...\n",
      " [1.         0.         0.         ... 0.5        0.46301525 0.        ]\n",
      " [1.         0.         0.         ... 0.875      0.47619048 0.        ]\n",
      " [1.         0.         0.         ... 0.921875   0.49689441 0.        ]]\n"
     ]
    }
   ],
   "source": [
    "print(w)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 683,
   "metadata": {},
   "outputs": [],
   "source": [
    "def costFunction(x,train_y,weights):\n",
    "    error = np.power(((x @ weights.T)-train_y),2)\n",
    "    temp = np.sum(error)/(2 * len(x))\n",
    "    return math.sqrt(temp/len(train_y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 684,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gradientDescent(x,y,epochs,learningRate,weights):\n",
    "    error = np.zeros(epochs)\n",
    "    for i in range(epochs):\n",
    "        weights = weights - (learningRate/len(x)) * np.sum(x * (x @ weights.T - y), axis=0)\n",
    "        error[i] = costFunction(x, y, weights)\n",
    "    return weights,error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 685,
   "metadata": {},
   "outputs": [],
   "source": [
    "y,ww = getMatrices(teF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 686,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1.         0.         0.         0.         1.         1.\n",
      " 0.         0.         1.         0.8125     0.56841709 1.        ]\n"
     ]
    }
   ],
   "source": [
    "print(y[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 687,
   "metadata": {},
   "outputs": [],
   "source": [
    "fw,error = gradientDescent(x,trT,epochs,learningRate,w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 688,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.01666938  0.01236114  0.00492266  0.0028857  -0.00350013 -0.18038337\n",
      "   0.19705275  0.00829951  0.00836987  0.25329431  0.20778187  0.00676421]]\n"
     ]
    }
   ],
   "source": [
    "print(fw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 689,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getAccuracy(finalWeights,x,y):\n",
    "    Y = np.dot(finalWeights,np.transpose(x))\n",
    "    Y = Y.ravel()\n",
    "    y = y.ravel()\n",
    "    print(len(Y))\n",
    "    print(len(y))\n",
    "    counter = 0\n",
    "    for i in range (0,len(Y)):\n",
    "            if(Y[i].round() == y[i].round()):\n",
    "                counter+=1\n",
    "    accuracy = (float((counter*100))/float(len(Y)))\n",
    "    return accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 690,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "133\n",
      "133\n",
      "96.2406015037594\n"
     ]
    }
   ],
   "source": [
    "accuracy = getAccuracy(fw,y,teT)\n",
    "print(accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 691,
   "metadata": {},
   "outputs": [],
   "source": [
    "def drawPlot(epochs,error):\n",
    "    dummy, axis = plt.subplots()  \n",
    "    axis.plot(np.arange(epochs), error, 'r')  \n",
    "    axis.set_xlabel('Epochs')  \n",
    "    axis.set_ylabel('Error')  \n",
    "    axis.set_title('Error vs. Epoch') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 692,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "drawPlot(epochs,error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}