In-Class Assignment 2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Step 1: Create three clusters of data: {xi^j}(i=1;1000) iid N(uj,sigj)(j=1,2,3)\n",
    "\n",
    "Color \n",
    "j = 1; Red; \n",
    "J = 2; Green;\n",
    "j = 3; Blue;\n",
    "\n",
    "u1 = [1 0]\n",
    "u2 = [-1 0]\n",
    "u3 = [0 =10]\n",
    "\n",
    "sig1 = u(lam1)u.t\n",
    "sig2 = u(lam2)u.t\n",
    "sig3 = I\n",
    "\n",
    "u = [[1/sqrt(2) -1/sqrt(2)] [1/sqrt(2) 1/sqrt(2)]]\n",
    "lam1 = [[4 0][0 1]]\n",
    "lam2 = [[1 0][0 5]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "u1 = [1,0]\n",
    "u2 = [-1,0]\n",
    "u3 = [0,-10]\n",
    "\n",
    "u = np.matrix([[1/np.sqrt(2),-1/np.sqrt(2)],[1/np.sqrt(2),1/np.sqrt(2)]])\n",
    "lam1 = np.matrix([[4,0],[0,1]])\n",
    "lam2 = np.matrix([[1,0],[0,5]])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 1.01360675  2.15315478]\n",
      " [ 2.36794274  3.34611408]\n",
      " [-0.47784012  0.64852498]\n",
      " ...\n",
      " [ 0.6883794   0.89831195]\n",
      " [ 1.88850721 -0.10399891]\n",
      " [ 2.15044973  0.28736936]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x194d2d8ec50>"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import scipy.stats \n",
    "training_data1 = scipy.stats.multivariate_normal(u1,u@(lam1)@u.T).rvs(1000)\n",
    "training_data2 = scipy.stats.multivariate_normal(u2,u@(lam2)@u.T).rvs(1000)\n",
    "training_data3 = scipy.stats.multivariate_normal(u3,np.identity(2)).rvs(1000)\n",
    "print(training_data1)\n",
    "plt.scatter(training_data1[:,0], training_data1[:,1], c='red')\n",
    "plt.scatter(training_data2[:,0], training_data2[:,1], c='green')\n",
    "plt.scatter(training_data3[:,0], training_data3[:,1], c='blue')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "K-Means Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1000, 2)\n",
      "(1000, 2)\n",
      "(1000, 2)\n",
      "(3000, 2)\n",
      "(3000, 2)\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "print(training_data1.shape)\n",
    "print(training_data2.shape)\n",
    "print(training_data3.shape)\n",
    "training_data = np.concatenate((training_data1,training_data2,training_data3))\n",
    "print(training_data.shape)\n",
    "random.shuffle(training_data)\n",
    "print(training_data.shape)\n",
    "cluster = np.zeros((3000))\n",
    "cluster[0:1000] = 0\n",
    "cluster[1000:2000] = 1\n",
    "cluster[2000:3000] = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.031731913400944584\n",
      "6000\n"
     ]
    }
   ],
   "source": [
    "xmean0 = np.mean(training_data[0:1000,0])\n",
    "xmean0\n",
    "ymean0 = np.mean(training_data[0:1000,1])\n",
    "xmean1 = np.mean(training_data[1000:2000,0])\n",
    "ymean1 = np.mean(training_data[1000:2000,1])\n",
    "print(ymean1)\n",
    "xmean2 = np.mean(training_data[2000:3000,0])\n",
    "ymean2 = np.mean(training_data[2000:3000,1])\n",
    "print(training_data.size)\n",
    "xarr = [0,1,2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "w = np.zeros((3));\n",
    "newcluster = np.zeros(3000)\n",
    "x = 0;\n",
    "while(not (np.array_equal(cluster,newcluster))):\n",
    "    cluster = newcluster;\n",
    "    xu = 0\n",
    "    yu = 0\n",
    "    zu = 0\n",
    "    xmean00 = 0\n",
    "    ymean00 = 0\n",
    "    xmean11 = 0\n",
    "    ymean11 = 0\n",
    "    xmean22 = 0\n",
    "    ymean22 = 0\n",
    "    \n",
    "    for j in range(3000):\n",
    "        w[0] = np.sqrt((xmean0 - training_data[j][0])**2 + (ymean0 - training_data[j][1])**2)\n",
    "        w[1] = np.sqrt((xmean1 - training_data[j][0])**2 + (ymean1 - training_data[j][1])**2)\n",
    "        w[2] = np.sqrt((xmean2 - training_data[j][0])**2 + (ymean2 - training_data[j][1])**2)\n",
    "        #print(np.argmin(w))\n",
    "        newcluster[j] = xarr[np.argmin(w)]\n",
    "    \n",
    "    for k in range(3000):\n",
    "        if(newcluster[q]==0):\n",
    "            xmean00 = training_data[k][0]\n",
    "            ymean00 = training_data[k][1]\n",
    "            xu = xu + 1\n",
    "        if(newcluster[q]==1):\n",
    "            xmean11 = training_data[k][0]\n",
    "            ymean11 = training_data[k][1]\n",
    "            yu = yu + 1\n",
    "        if(newcluster[q]==2):\n",
    "            xmean22 = training_data[k][0]\n",
    "            ymean22 = training_data[k][1]\n",
    "            zu = zu + 1\n",
    "    if(xu != 0):\n",
    "        xmean0 = xmean00/xu\n",
    "        ymean0 = ymean00/xu\n",
    "    if(yu != 0):\n",
    "        xmean1 = xmean11/yu\n",
    "        ymean1 = ymean11/yu\n",
    "    if(zu != 0):\n",
    "        xmean2 = xmean22/zu\n",
    "        ymean2 = ymean22/zu\n",
    "    \n",
    "# newcluster = np.around(newcluster,decimals = 1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "for q in range (3000): \n",
    "        if (newcluster[q]==0):\n",
    "            plt.scatter(training_data[q,0],training_data[q,1],color= 'red')\n",
    "        if (newcluster[q]==1):\n",
    "            plt.scatter(training_data[q,0],training_data[q,1],color= 'blue')\n",
    "        if (newcluster[q]==2):\n",
    "            plt.scatter(training_data[q,0],training_data[q,1],color= 'green')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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
}