Classification-Toy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification Problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Module imports and constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from src.RBF import RBFNet\n",
    "from src.ES import ES\n",
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import preprocessing\n",
    "from sklearn import datasets\n",
    "\n",
    "MARKERS = ['s', 'o', 'd', 'X', 'P', '8']\n",
    "COLORS = ['r', 'g', 'b', 'c', 'm', 'y']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## First Example: Basic"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialization and plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "points = np.loadtxt('data1.txt', delimiter=',')\n",
    "n = points.shape[1] - 1 #features\n",
    "m = points.shape[0] # rows\n",
    "X = points[:, :n]\n",
    "y = points[:, n:]\n",
    "for i in range(m): \n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ES initialization and training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "start training\n",
      "gen\tnevals\tavg    \tstd    \tmin    \tmax    \n",
      "0  \t100   \t83.2761\t8.04083\t54.0589\t87.8138\n",
      "1  \t150   \t71.4219\t9.89525\t49.9717\t87.6928\n",
      "2  \t150   \t63.4644\t12.551 \t29.1514\t87.5029\n",
      "3  \t150   \t52.8834\t13.0624\t20.7561\t87.4592\n",
      "4  \t150   \t43.7065\t13.4951\t16.5884\t70.8133\n",
      "5  \t150   \t31.9565\t12.5038\t4.03141\t70.8033\n",
      "6  \t150   \t18.0086\t9.12427\t3.94425\t37.4943\n",
      "7  \t150   \t11.4302\t8.79663\t3.90067\t37.5372\n",
      "8  \t150   \t7.36758\t5.77599\t3.89099\t24.8148\n",
      "9  \t150   \t5.67335\t3.4643 \t3.85795\t20.7241\n",
      "10 \t150   \t4.98659\t2.74471\t3.85381\t20.6652\n",
      "11 \t150   \t4.20581\t1.3679 \t3.84901\t12.3504\n",
      "12 \t150   \t4.15718\t1.54751\t3.85057\t16.4606\n",
      "13 \t150   \t4.05644\t1.01754\t3.85046\t12.3   \n",
      "14 \t150   \t3.88441\t0.0262762\t3.84611\t3.96625\n",
      "15 \t150   \t4.00855\t0.932742 \t3.84674\t12.2887\n",
      "16 \t150   \t3.91141\t0.424144 \t3.84671\t8.12528\n",
      "17 \t150   \t3.86372\t0.0144089\t3.845  \t3.90869\n",
      "18 \t150   \t3.90071\t0.42513  \t3.84508\t8.12867\n",
      "19 \t150   \t3.85197\t0.00684902\t3.84537\t3.89198\n",
      "20 \t150   \t3.84907\t0.0035498 \t3.84464\t3.86953\n",
      "21 \t150   \t3.84728\t0.00111031\t3.84523\t3.85115\n",
      "22 \t150   \t3.84673\t0.000815442\t3.84433\t3.85062\n",
      "23 \t150   \t3.84609\t0.000710742\t3.84433\t3.84752\n",
      "24 \t150   \t3.84572\t0.000613652\t3.84436\t3.84722\n",
      "25 \t150   \t3.84541\t0.000410252\t3.84439\t3.84621\n",
      "26 \t150   \t3.84515\t0.000346973\t3.84432\t3.8461 \n",
      "27 \t150   \t3.8449 \t0.000277858\t3.84425\t3.84562\n",
      "28 \t150   \t3.84472\t0.000189697\t3.84428\t3.84517\n",
      "29 \t150   \t3.84457\t0.000164562\t3.84406\t3.84503\n",
      "30 \t150   \t3.84447\t0.000137273\t3.84413\t3.84482\n",
      "31 \t150   \t3.84439\t9.33448e-05\t3.84423\t3.84461\n",
      "32 \t150   \t3.84431\t9.69432e-05\t3.84403\t3.84455\n",
      "33 \t150   \t3.84424\t7.682e-05  \t3.84407\t3.84441\n",
      "34 \t150   \t3.84417\t7.27084e-05\t3.84402\t3.84434\n",
      "35 \t150   \t3.84414\t5.06371e-05\t3.84398\t3.84426\n",
      "36 \t150   \t3.84411\t4.46894e-05\t3.84401\t3.84424\n",
      "37 \t150   \t3.84408\t3.54838e-05\t3.84399\t3.84415\n",
      "38 \t150   \t3.84406\t3.3614e-05 \t3.84399\t3.84414\n",
      "39 \t150   \t3.84404\t2.53544e-05\t3.84398\t3.8441 \n",
      "40 \t150   \t3.84403\t1.95706e-05\t3.84397\t3.84408\n",
      "best individual's error: [[3.84397394]]\n"
     ]
    }
   ],
   "source": [
    "es = ES(X=X, y=y, MU=100, LAMBDA=150, ngen=40, circles=6, problem='classification')\n",
    "my_rbf = es.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
    "y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
    "plt.xlim(x_min, x_max)\n",
    "plt.ylim(y_min, y_max)\n",
    "x_range = np.arange(x_min, x_max, 0.1)\n",
    "y_range = np.arange(y_min, y_max, 0.1)\n",
    "xx, yy = np.meshgrid(x_range, y_range)\n",
    "cmap = plt.get_cmap('Paired')\n",
    "zz = np.zeros(xx.shape)\n",
    "\n",
    "for i in range(zz.shape[0]):\n",
    "    for j in range(zz.shape[1]):\n",
    "        x_vector = np.array([xx[i][j], yy[i][j]])\n",
    "        network_answer = my_rbf.apply_network(x_vector)\n",
    "        cls = np.argmax(network_answer)\n",
    "        zz[i][j] = int(cls)\n",
    "\n",
    "plt.pcolormesh(xx, yy, zz, cmap=cmap)\n",
    "\n",
    "for i in range(m): \n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Second Example"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialization and plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "points = np.loadtxt('data2.txt', delimiter=',')\n",
    "np.random.shuffle(points)\n",
    "n = points.shape[1] - 1 #features\n",
    "m = points.shape[0] # rows\n",
    "X = points[:, :n]\n",
    "X = preprocessing.scale(X)\n",
    "y = points[:, n:]\n",
    "\n",
    "for i in range(m): \n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ES initialization and training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "start training\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'FitnessMin' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'Individual' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'Strategy' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gen\tnevals\tavg    \tstd     \tmin  \tmax    \n",
      "0  \t100   \t50.4487\t0.951895\t42.95\t53.9592\n",
      "1  \t150   \t49.9393\t2.20523 \t38.6759\t50.6006\n",
      "2  \t150   \t49.3703\t3.36141 \t33.5784\t50.5982\n",
      "3  \t150   \t48.5091\t4.20638 \t31.0061\t50.5901\n",
      "4  \t150   \t46.3889\t6.32654 \t25.0365\t55.6699\n",
      "5  \t150   \t41.9386\t7.0712  \t24.1802\t52.2945\n",
      "6  \t150   \t38.0701\t8.46157 \t19.1161\t53.1208\n",
      "7  \t150   \t36.16  \t9.38159 \t17.3639\t55.6668\n",
      "8  \t150   \t29.5604\t8.42632 \t16.4908\t50.589 \n",
      "9  \t150   \t26.4719\t8.47311 \t16.4908\t47.1843\n",
      "10 \t150   \t22.376 \t5.59137 \t14.7964\t39.5486\n",
      "11 \t150   \t20.095 \t4.24296 \t14.7901\t35.2814\n",
      "12 \t150   \t19.1751\t4.60487 \t13.9647\t36.1463\n",
      "13 \t150   \t17.3738\t3.3281  \t14.7857\t32.7242\n",
      "14 \t150   \t16.3257\t1.37946 \t14.7905\t21.5986\n",
      "15 \t150   \t15.8999\t1.07027 \t13.9557\t19.9004\n",
      "16 \t150   \t15.7124\t0.918315\t13.9592\t20.7573\n",
      "17 \t150   \t15.6266\t0.873458\t13.9565\t20.7557\n",
      "18 \t150   \t15.5929\t0.707287\t13.9536\t18.2021\n",
      "19 \t150   \t15.7288\t0.798064\t13.9527\t19.0473\n",
      "20 \t150   \t15.4669\t0.612944\t13.9525\t17.3462\n",
      "21 \t150   \t15.2724\t0.564155\t13.954 \t16.489 \n",
      "22 \t150   \t15.0866\t0.612203\t13.9509\t16.4915\n",
      "23 \t150   \t15.0022\t0.508753\t13.9513\t15.6469\n",
      "24 \t150   \t14.8416\t0.324579\t13.9514\t15.6454\n",
      "25 \t150   \t14.7314\t0.285302\t13.9511\t15.6444\n",
      "26 \t150   \t14.7057\t0.264719\t13.9519\t14.8007\n",
      "27 \t150   \t14.7564\t0.184567\t13.9515\t14.8008\n",
      "28 \t150   \t14.7392\t0.216014\t13.9514\t14.7996\n",
      "29 \t150   \t14.7813\t0.118559\t13.9514\t14.7991\n",
      "30 \t150   \t14.7895\t0.084219\t13.9516\t14.799 \n",
      "31 \t150   \t14.7894\t0.0842005\t13.9516\t14.7985\n",
      "32 \t150   \t14.7978\t0.000231529\t14.797 \t14.7983\n",
      "33 \t150   \t14.7976\t0.00019543 \t14.7972\t14.7982\n",
      "34 \t150   \t14.7975\t0.000126449\t14.7971\t14.7978\n",
      "35 \t150   \t14.7974\t7.20334e-05\t14.7972\t14.7976\n",
      "36 \t150   \t14.7974\t6.19154e-05\t14.7972\t14.7975\n",
      "37 \t150   \t14.7973\t5.09803e-05\t14.7972\t14.7975\n",
      "38 \t150   \t14.7973\t3.49259e-05\t14.7972\t14.7974\n",
      "39 \t150   \t14.7973\t3.31891e-05\t14.7972\t14.7974\n",
      "40 \t150   \t14.7973\t2.80364e-05\t14.7972\t14.7973\n",
      "best individual's error: [[13.95093384]]\n"
     ]
    }
   ],
   "source": [
    "es = ES(X=X, y=y, MU=100, LAMBDA=150, ngen=40, circles=6, problem='classification')\n",
    "my_rbf = es.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "x_min, x_max = X[:, 0].min() - 0.05, X[:, 0].max() + 0.05\n",
    "y_min, y_max = X[:, 1].min() - 0.05, X[:, 1].max() + 0.05\n",
    "plt.xlim(x_min, x_max)\n",
    "plt.ylim(y_min, y_max)\n",
    "x_range = np.arange(x_min, x_max, 0.05)\n",
    "y_range = np.arange(y_min, y_max, 0.05)\n",
    "xx, yy = np.meshgrid(x_range, y_range)\n",
    "cmap = plt.get_cmap('PiYG')\n",
    "zz = np.zeros(xx.shape)\n",
    "\n",
    "for i in range(zz.shape[0]):\n",
    "    for j in range(zz.shape[1]):\n",
    "        x_vector = np.array([xx[i][j], yy[i][j]])\n",
    "        network_answer = my_rbf.apply_network(x_vector)\n",
    "        cls = np.argmax(network_answer)\n",
    "        zz[i][j] = int(cls)\n",
    "\n",
    "plt.pcolormesh(xx, yy, zz, cmap=cmap)\n",
    "\n",
    "for i in range(m): \n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: %87.28813559322033\n"
     ]
    }
   ],
   "source": [
    "dataset_size = X.shape[0]\n",
    "correct = 0\n",
    "for i in range(dataset_size): \n",
    "    actual_class = int(y[i][0])\n",
    "    predicted_class = int(np.argmax(my_rbf.apply_network(X[i])))\n",
    "    if actual_class == predicted_class:\n",
    "        correct += 1\n",
    "print('Training Accuracy: %{0}'.format(100 * correct / dataset_size))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Third Example: The Spiral Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialization and plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "N = 100 # number of points per class\n",
    "D = 2 # dimensionality\n",
    "K = 3 # number of classes\n",
    "X = np.zeros((N*K,D)) # data matrix (each row = single example)\n",
    "y = np.zeros(N*K, dtype='uint8') # class labels\n",
    "for j in range(K):\n",
    "  ix = range(N*j,N*(j+1))\n",
    "  r = np.linspace(0.0,1,N) # radius\n",
    "  t = np.linspace(j*4,(j+1)*4,N) + np.random.randn(N)*0.2 # theta\n",
    "  X[ix] = np.c_[r*np.sin(t), r*np.cos(t)]\n",
    "  y[ix] = j\n",
    "# Plot the data\n",
    "y = y.reshape(y.shape[0], 1)\n",
    "\n",
    "for i in range(X.shape[0]):\n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ES Initialization and training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "start training\n",
      "gen\tnevals\tavg    \tstd    \tmin    \tmax    \n",
      "0  \t100   \t63.1788\t7.42253\t44.5187\t69.8836\n",
      "1  \t150   \t56.5127\t8.25541\t39.4448\t68.9315\n",
      "2  \t150   \t52.2396\t8.0478 \t37.1146\t68.9463\n",
      "3  \t150   \t47.8901\t5.331  \t36.7773\t59.1913\n",
      "4  \t150   \t45.1971\t4.24796\t35.4351\t58.8767\n",
      "5  \t150   \t42.9389\t5.86597\t25.6449\t50.1329\n",
      "6  \t150   \t42.9969\t4.48379\t29.0477\t49.4377\n",
      "7  \t150   \t41.7865\t5.83083\t25.3028\t49.1157\n",
      "8  \t150   \t40.9838\t5.64675\t22.2983\t47.4292\n",
      "9  \t150   \t35.3837\t7.86951\t20.9847\t48.0875\n",
      "10 \t150   \t30.0674\t7.07267\t14.5709\t45.7323\n",
      "11 \t150   \t26.8732\t6.50678\t14.5615\t43.055 \n",
      "12 \t150   \t24.6341\t5.2233 \t13.2343\t45.0783\n",
      "13 \t150   \t23.8173\t4.6304 \t14.2509\t40.7236\n",
      "14 \t150   \t22.2222\t3.83673\t15.2542\t38.3923\n",
      "15 \t150   \t22.0354\t3.3755 \t14.2511\t32.0048\n",
      "16 \t150   \t20.7462\t2.7834 \t14.5815\t24.9953\n",
      "17 \t150   \t21.2614\t2.55541\t14.5868\t25.3115\n",
      "18 \t150   \t21.1278\t2.35591\t14.9347\t24.9783\n",
      "19 \t150   \t20.713 \t2.31421\t14.9299\t24.6459\n",
      "20 \t150   \t19.5743\t2.05884\t14.2587\t23.6469\n",
      "21 \t150   \t19.4815\t1.97651\t13.9223\t23.9731\n",
      "22 \t150   \t19.1794\t1.3978 \t15.9463\t22.6289\n",
      "23 \t150   \t18.5863\t1.28398\t15.255 \t21.2762\n",
      "24 \t150   \t18.1233\t1.56228\t13.9081\t22.6245\n",
      "25 \t150   \t18.1129\t1.39782\t13.2393\t21.2858\n",
      "26 \t150   \t17.8984\t1.49588\t13.5738\t20.6057\n",
      "27 \t150   \t17.6603\t1.31531\t14.9157\t20.2796\n",
      "28 \t150   \t17.5431\t0.988468\t15.2498\t20.3079\n",
      "29 \t150   \t17.2348\t0.671375\t15.5926\t18.6007\n",
      "30 \t150   \t16.997 \t0.635997\t14.9142\t18.2654\n",
      "31 \t150   \t16.7895\t0.396334\t14.9187\t17.9297\n",
      "32 \t150   \t16.749 \t0.209091\t15.9209\t17.2583\n",
      "33 \t150   \t16.6782\t0.204493\t15.9195\t16.9285\n",
      "34 \t150   \t16.6045\t0.115122\t16.2553\t16.9271\n",
      "35 \t150   \t16.5839\t0.0816096\t16.2546\t16.9256\n",
      "36 \t150   \t16.5868\t0.0578581\t16.2549\t16.9233\n",
      "37 \t150   \t16.5798\t0.0570661\t16.2553\t16.5915\n",
      "38 \t150   \t16.5759\t0.065559 \t16.2543\t16.5914\n",
      "39 \t150   \t16.5689\t0.0792908\t16.2542\t16.59  \n",
      "40 \t150   \t16.5518\t0.123948 \t15.9208\t16.5895\n",
      "best individual's error: [[13.23426832]]\n"
     ]
    }
   ],
   "source": [
    "es = ES(X=X, y=y, MU=100, LAMBDA=150, ngen=40, circles=8, problem='classification')\n",
    "my_rbf = es.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "x_min, x_max = X[:, 0].min() - 0.05, X[:, 0].max() + 0.05\n",
    "y_min, y_max = X[:, 1].min() - 0.05, X[:, 1].max() + 0.05\n",
    "plt.xlim(x_min, x_max)\n",
    "plt.ylim(y_min, y_max)\n",
    "x_range = np.arange(x_min, x_max, 0.05)\n",
    "y_range = np.arange(y_min, y_max, 0.05)\n",
    "xx, yy = np.meshgrid(x_range, y_range)\n",
    "cmap = plt.get_cmap('Paired')\n",
    "zz = np.zeros(xx.shape)\n",
    "\n",
    "for i in range(zz.shape[0]):\n",
    "    for j in range(zz.shape[1]):\n",
    "        x_vector = np.array([xx[i][j], yy[i][j]])\n",
    "        network_answer = my_rbf.apply_network(x_vector)\n",
    "        cls = np.argmax(network_answer)\n",
    "        zz[i][j] = int(cls)\n",
    "\n",
    "plt.pcolormesh(xx, yy, zz, cmap=cmap)\n",
    "\n",
    "for i in range(X.shape[0]):\n",
    "    a = X[i][0]\n",
    "    b = X[i][1]\n",
    "    cls = int(y[i][0])\n",
    "    plt.scatter(a, b, marker=MARKERS[cls], color=COLORS[cls])\n",
    "    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: %88.33333333333333\n"
     ]
    }
   ],
   "source": [
    "dataset_size = X.shape[0]\n",
    "correct = 0\n",
    "for i in range(dataset_size): \n",
    "    actual_class = int(y[i][0])\n",
    "    predicted_class = int(np.argmax(my_rbf.apply_network(X[i])))\n",
    "    if actual_class == predicted_class:\n",
    "        correct += 1\n",
    "print('Training Accuracy: %{0}'.format(100 * correct / dataset_size))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fourth Example: The Iris Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Initialization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = datasets.load_iris(True)\n",
    "m, n = X.shape\n",
    "y = y.reshape(m, 1)\n",
    "X = preprocessing.scale(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ES initialization and training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "start training\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'FitnessMin' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'Individual' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\creator.py:141: RuntimeWarning: A class named 'Strategy' has already been created and it will be overwritten. Consider deleting previous creation of that class or rename it.\n",
      "  RuntimeWarning)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\numpy\\linalg\\linalg.py:1943: RuntimeWarning: overflow encountered in true_divide\n",
      "  s = divide(1, s, where=large, out=s)\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\numpy\\linalg\\linalg.py:1946: RuntimeWarning: invalid value encountered in multiply\n",
      "  res = matmul(transpose(vt), multiply(s[..., newaxis], transpose(u)))\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\numpy\\linalg\\linalg.py:1946: RuntimeWarning: invalid value encountered in matmul\n",
      "  res = matmul(transpose(vt), multiply(s[..., newaxis], transpose(u)))\n",
      "c:\\users\\mohammad\\appdata\\local\\programs\\python\\python36-32\\lib\\site-packages\\deap\\base.py:236: RuntimeWarning: invalid value encountered in less_equal\n",
      "  return self.wvalues <= other.wvalues\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gen\tnevals\tavg\tstd\tmin\tmax\n",
      "0  \t100   \tnan\tnan\tnan\tnan\n",
      "1  \t150   \t65.984\t8.27327\t32.7227\t68.7868\n",
      "2  \t150   \t58.45 \t15.5861\t10.7377\t68.7701\n",
      "3  \t150   \t55.0676\t14.0407\t24.7443\t68.857 \n",
      "4  \t150   \t53.6402\t15.4387\t19.9622\t68.822 \n",
      "5  \t150   \t43.4698\t13.2618\t17.3493\t68.7954\n",
      "6  \t150   \t40.1562\t13.3005\t5.9921 \t68.7686\n",
      "7  \t150   \t35.121 \t12.4454\t10.6337\t68.7204\n",
      "8  \t150   \t26.1387\t10.8783\t6.49595\t68.6655\n",
      "9  \t150   \t21.1833\t9.53274\t7.20666\t68.6499\n",
      "10 \t150   \t17.0872\t7.20225\t5.83054\t35.3435\n",
      "11 \t150   \t14.1154\t5.35801\t5.23558\t29.365 \n",
      "12 \t150   \t12.5977\t4.38752\t5.92339\t27.9762\n",
      "13 \t150   \t12.05  \t4.56506\t5.17293\t26.6227\n",
      "14 \t150   \t10.3429\t3.66015\t5.1839 \t23.3093\n",
      "15 \t150   \t8.40545\t2.89355\t5.15935\t18.581 \n",
      "16 \t150   \t6.91196\t1.52741\t4.49059\t13.8933\n",
      "17 \t150   \t6.11596\t0.85229\t4.48891\t9.18516\n",
      "18 \t150   \t5.85169\t0.629695\t4.50073\t7.84001\n",
      "19 \t150   \t5.5167 \t0.447138\t4.49199\t6.50171\n",
      "20 \t150   \t5.32124\t0.328747\t4.49742\t6.49945\n",
      "21 \t150   \t5.16002\t0.326358\t4.48952\t6.49482\n",
      "22 \t150   \t5.08518\t0.24663 \t4.49178\t5.82328\n",
      "23 \t150   \t5.0178 \t0.27058 \t4.49042\t5.16217\n",
      "24 \t150   \t5.01701\t0.270243\t4.49065\t5.15928\n",
      "25 \t150   \t5.0296 \t0.26008 \t4.48927\t5.15985\n",
      "26 \t150   \t5.03587\t0.25479 \t4.48817\t5.15865\n",
      "27 \t150   \t5.02856\t0.260438\t4.48884\t5.15714\n",
      "28 \t150   \t5.06788\t0.223425\t4.48793\t5.15795\n",
      "29 \t150   \t5.08742\t0.199677\t4.48744\t5.15594\n",
      "30 \t150   \t5.1272 \t0.130242\t4.48915\t5.15439\n",
      "31 \t150   \t5.15363\t0.000298492\t5.15302\t5.15477\n",
      "32 \t150   \t5.15342\t0.000205232\t5.153  \t5.15391\n",
      "33 \t150   \t5.15331\t0.000226187\t5.15277\t5.1538 \n",
      "34 \t150   \t5.15323\t0.000143504\t5.15277\t5.15353\n",
      "35 \t150   \t5.1531 \t0.000138596\t5.1528 \t5.15332\n",
      "36 \t150   \t5.15301\t0.000143859\t5.15264\t5.15332\n",
      "37 \t150   \t5.15292\t0.000110016\t5.15262\t5.15317\n",
      "38 \t150   \t5.15285\t9.02506e-05\t5.15265\t5.15308\n",
      "39 \t150   \t5.1528 \t8.15973e-05\t5.15263\t5.15302\n",
      "40 \t150   \t5.15276\t5.74091e-05\t5.15262\t5.15292\n",
      "41 \t150   \t5.15272\t5.08049e-05\t5.1526 \t5.15289\n",
      "42 \t150   \t5.15269\t4.73343e-05\t5.15258\t5.15282\n",
      "43 \t150   \t5.15265\t3.55529e-05\t5.15257\t5.15273\n",
      "44 \t150   \t5.15263\t2.74382e-05\t5.15256\t5.1527 \n",
      "45 \t150   \t5.15262\t2.03747e-05\t5.15256\t5.15266\n",
      "46 \t150   \t5.15261\t1.69247e-05\t5.15257\t5.15265\n",
      "47 \t150   \t5.1526 \t1.11459e-05\t5.15258\t5.15264\n",
      "48 \t150   \t5.1526 \t9.12566e-06\t5.15257\t5.15262\n",
      "49 \t150   \t5.15259\t6.63198e-06\t5.15257\t5.15262\n",
      "50 \t150   \t5.15259\t4.12379e-06\t5.15258\t5.1526 \n",
      "best individual's error: [[4.48744141]]\n"
     ]
    }
   ],
   "source": [
    "es = ES(X=X, y=y, MU=100, LAMBDA=150, ngen=50, circles=5, problem='classification')\n",
    "my_rbf = es.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: %97.33333333333333\n"
     ]
    }
   ],
   "source": [
    "dataset_size = X.shape[0]\n",
    "correct = 0\n",
    "for i in range(dataset_size): \n",
    "    actual_class = int(y[i][0])\n",
    "    predicted_class = int(np.argmax(my_rbf.apply_network(X[i])))\n",
    "    if actual_class == predicted_class:\n",
    "        correct += 1\n",
    "print('Training Accuracy: %{0}'.format(100 * correct / dataset_size))"
   ]
  }
 ],
 "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
}