SupportVectorMachinesReview.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "library(plyr)\n",
    "library(ggplot2)\n",
    "library(e1071)\n",
    "library(caret) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Quick example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th scope=col>X</th><th scope=col>Y</th><th scope=col>class</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><td>4.915895</td><td>3.312657</td><td>1       </td></tr>\n",
       "\t<tr><td>2.203087</td><td>2.762946</td><td>1       </td></tr>\n",
       "\t<tr><td>2.376347</td><td>2.273357</td><td>1       </td></tr>\n",
       "\t<tr><td>4.366696</td><td>2.682983</td><td>1       </td></tr>\n",
       "\t<tr><td>2.079594</td><td>2.194263</td><td>1       </td></tr>\n",
       "\t<tr><td>4.141328</td><td>3.377833</td><td>1       </td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|lll}\n",
       " X & Y & class\\\\\n",
       "\\hline\n",
       "\t 4.915895 & 3.312657 & 1       \\\\\n",
       "\t 2.203087 & 2.762946 & 1       \\\\\n",
       "\t 2.376347 & 2.273357 & 1       \\\\\n",
       "\t 4.366696 & 2.682983 & 1       \\\\\n",
       "\t 2.079594 & 2.194263 & 1       \\\\\n",
       "\t 4.141328 & 3.377833 & 1       \\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "X | Y | class | \n",
       "|---|---|---|---|---|---|\n",
       "| 4.915895 | 3.312657 | 1        | \n",
       "| 2.203087 | 2.762946 | 1        | \n",
       "| 2.376347 | 2.273357 | 1        | \n",
       "| 4.366696 | 2.682983 | 1        | \n",
       "| 2.079594 | 2.194263 | 1        | \n",
       "| 4.141328 | 3.377833 | 1        | \n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  X        Y        class\n",
       "1 4.915895 3.312657 1    \n",
       "2 2.203087 2.762946 1    \n",
       "3 2.376347 2.273357 1    \n",
       "4 4.366696 2.682983 1    \n",
       "5 2.079594 2.194263 1    \n",
       "6 4.141328 3.377833 1    "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "points_data1 <- read.csv(\"datasets/data1.csv\", header = TRUE)\n",
    "head(points_data1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th scope=col>class</th><th scope=col>freq</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><td>-1</td><td>75</td></tr>\n",
       "\t<tr><td> 1</td><td>75</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|ll}\n",
       " class & freq\\\\\n",
       "\\hline\n",
       "\t -1 & 75\\\\\n",
       "\t  1 & 75\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "class | freq | \n",
       "|---|---|\n",
       "| -1 | 75 | \n",
       "|  1 | 75 | \n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  class freq\n",
       "1 -1    75  \n",
       "2  1    75  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "count(points_data1, \"class\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "points_data1$class <- factor(points_data1$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {},
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAAP1BMVEUAAAAAv8QzMzNNTU1o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\nEBKAAf8Pcqjsd7oOZ5cAAAAASUVORK5CYII=",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot1 <- ggplot(points_data1, aes(X, Y)) + \n",
    "            geom_point(aes(color = factor(class), shape = factor(class)))\n",
    "\n",
    "plot1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = points_data1, kernel = \"linear\", \n",
       "    cost = 10, scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  linear \n",
       "       cost:  10 \n",
       "      gamma:  0.5 \n",
       "\n",
       "Number of Support Vectors:  3\n",
       "\n",
       " ( 2 1 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_model1 <- svm(class ~. ,\n",
    "                     data = points_data1,\n",
    "                     kernel = \"linear\",\n",
    "                     cost = 10, \n",
    "                     scale = FALSE)\n",
    "\n",
    "summary(svm_model1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAA/FBMVEUAAAASJSUUKioVKioc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\nRSJEAYpEiAIUiRAF/j/KFm6ZljQT0QAAAABJRU5ErkJggg==",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model1, points_data1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We trained the SVM with a linear kernel; and it learned a linear decision boundary. The \"X\" marks in the above plot are the data points that serve as support vectors. There are three of them. Let's compute the confusion matrix, and as expected, we'll get perfect accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred <- predict(svm_model1, points_data1[, -3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 75  0\n",
       "        1   0 75\n",
       "                                     \n",
       "               Accuracy : 1          \n",
       "                 95% CI : (0.9757, 1)\n",
       "    No Information Rate : 0.5        \n",
       "    P-Value [Acc > NIR] : < 2.2e-16  \n",
       "                                     \n",
       "                  Kappa : 1          \n",
       " Mcnemar's Test P-Value : NA         \n",
       "                                     \n",
       "            Sensitivity : 1.0        \n",
       "            Specificity : 1.0        \n",
       "         Pos Pred Value : 1.0        \n",
       "         Neg Pred Value : 1.0        \n",
       "             Prevalence : 0.5        \n",
       "         Detection Rate : 0.5        \n",
       "   Detection Prevalence : 0.5        \n",
       "      Balanced Accuracy : 1.0        \n",
       "                                     \n",
       "       'Positive' Class : -1         \n",
       "                                     "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(pred, points_data1$class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A bit more complicated example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "points_data <- read.csv(\"datasets/data3.csv\", header = TRUE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th scope=col>X</th><th scope=col>Y</th><th scope=col>class</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><td>1.2930150</td><td>0.7039713</td><td>1        </td></tr>\n",
       "\t<tr><td>0.9789579</td><td>0.5897531</td><td>1        </td></tr>\n",
       "\t<tr><td>0.6907754</td><td>1.5024883</td><td>1        </td></tr>\n",
       "\t<tr><td>1.5548983</td><td>1.2567174</td><td>1        </td></tr>\n",
       "\t<tr><td>1.1917170</td><td>0.7128628</td><td>1        </td></tr>\n",
       "\t<tr><td>1.0199767</td><td>0.9425957</td><td>1        </td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|lll}\n",
       " X & Y & class\\\\\n",
       "\\hline\n",
       "\t 1.2930150 & 0.7039713 & 1        \\\\\n",
       "\t 0.9789579 & 0.5897531 & 1        \\\\\n",
       "\t 0.6907754 & 1.5024883 & 1        \\\\\n",
       "\t 1.5548983 & 1.2567174 & 1        \\\\\n",
       "\t 1.1917170 & 0.7128628 & 1        \\\\\n",
       "\t 1.0199767 & 0.9425957 & 1        \\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "X | Y | class | \n",
       "|---|---|---|---|---|---|\n",
       "| 1.2930150 | 0.7039713 | 1         | \n",
       "| 0.9789579 | 0.5897531 | 1         | \n",
       "| 0.6907754 | 1.5024883 | 1         | \n",
       "| 1.5548983 | 1.2567174 | 1         | \n",
       "| 1.1917170 | 0.7128628 | 1         | \n",
       "| 1.0199767 | 0.9425957 | 1         | \n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  X         Y         class\n",
       "1 1.2930150 0.7039713 1    \n",
       "2 0.9789579 0.5897531 1    \n",
       "3 0.6907754 1.5024883 1    \n",
       "4 1.5548983 1.2567174 1    \n",
       "5 1.1917170 0.7128628 1    \n",
       "6 1.0199767 0.9425957 1    "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "head(points_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th scope=col>class</th><th scope=col>freq</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><td>-1 </td><td>125</td></tr>\n",
       "\t<tr><td> 1 </td><td>125</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|ll}\n",
       " class & freq\\\\\n",
       "\\hline\n",
       "\t -1  & 125\\\\\n",
       "\t  1  & 125\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "class | freq | \n",
       "|---|---|\n",
       "| -1  | 125 | \n",
       "|  1  | 125 | \n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  class freq\n",
       "1 -1    125 \n",
       "2  1    125 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "count(points_data, \"class\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "points_data$class <- factor(points_data$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {},
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAAP1BMVEUAAAAAv8QzMzNNTU1o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\nAAAASUVORK5CYII=",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot1 <- ggplot(points_data, aes(X, Y)) + \n",
    "            geom_point(aes(color = factor(class), shape = factor(class)))\n",
    "\n",
    "plot1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = points_data, kernel = \"radial\", cost = 10, \n",
       "    scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  radial \n",
       "       cost:  10 \n",
       "      gamma:  0.5 \n",
       "\n",
       "Number of Support Vectors:  37\n",
       "\n",
       " ( 20 17 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_model1 <- svm(class ~. ,\n",
    "                     data = points_data,\n",
    "                     kernel = \"radial\",\n",
    "                     cost = 10, \n",
    "                     scale = FALSE)\n",
    "\n",
    "summary(svm_model1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAB+FBMVEUAAAALFxcPHx8SJSUU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\ndPSGowAAAABJRU5ErkJggg==",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model1, points_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred <- predict(svm_model1, points_data[, -3]) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction  -1   1\n",
       "        -1 119   8\n",
       "        1    6 117\n",
       "                                         \n",
       "               Accuracy : 0.944          \n",
       "                 95% CI : (0.9078, 0.969)\n",
       "    No Information Rate : 0.5            \n",
       "    P-Value [Acc > NIR] : <2e-16         \n",
       "                                         \n",
       "                  Kappa : 0.888          \n",
       " Mcnemar's Test P-Value : 0.7893         \n",
       "                                         \n",
       "            Sensitivity : 0.9520         \n",
       "            Specificity : 0.9360         \n",
       "         Pos Pred Value : 0.9370         \n",
       "         Neg Pred Value : 0.9512         \n",
       "             Prevalence : 0.5000         \n",
       "         Detection Rate : 0.4760         \n",
       "   Detection Prevalence : 0.5080         \n",
       "      Balanced Accuracy : 0.9440         \n",
       "                                         \n",
       "       'Positive' Class : -1             \n",
       "                                         "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(pred, points_data$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is a potential problem here. We trained and tested our model with the same exact data set. This can cause memorization; the model does not learn a decision boundary, it memorizes a boundary for this particular data set. We don't know how it will perform on the new, unseen observations. One of the most important aspects of learning algorithms is their ability to generalize; that is, to learn decision boundaries that are generalized enough to do well on unseen data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Split test\n",
    "\n",
    "Split test is simply splitting the data into training and testing sets; \n",
    "we can use 65% of the data for training, and the rest for testing "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "split <- 0.65"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>1</li>\n",
       "\t<li>2</li>\n",
       "\t<li>3</li>\n",
       "\t<li>8</li>\n",
       "\t<li>9</li>\n",
       "\t<li>10</li>\n",
       "\t<li>11</li>\n",
       "\t<li>12</li>\n",
       "\t<li>13</li>\n",
       "\t<li>15</li>\n",
       "\t<li>17</li>\n",
       "\t<li>18</li>\n",
       "\t<li>19</li>\n",
       "\t<li>20</li>\n",
       "\t<li>22</li>\n",
       "\t<li>24</li>\n",
       "\t<li>25</li>\n",
       "\t<li>26</li>\n",
       "\t<li>28</li>\n",
       "\t<li>29</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 1\n",
       "\\item 2\n",
       "\\item 3\n",
       "\\item 8\n",
       "\\item 9\n",
       "\\item 10\n",
       "\\item 11\n",
       "\\item 12\n",
       "\\item 13\n",
       "\\item 15\n",
       "\\item 17\n",
       "\\item 18\n",
       "\\item 19\n",
       "\\item 20\n",
       "\\item 22\n",
       "\\item 24\n",
       "\\item 25\n",
       "\\item 26\n",
       "\\item 28\n",
       "\\item 29\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 1\n",
       "2. 2\n",
       "3. 3\n",
       "4. 8\n",
       "5. 9\n",
       "6. 10\n",
       "7. 11\n",
       "8. 12\n",
       "9. 13\n",
       "10. 15\n",
       "11. 17\n",
       "12. 18\n",
       "13. 19\n",
       "14. 20\n",
       "15. 22\n",
       "16. 24\n",
       "17. 25\n",
       "18. 26\n",
       "19. 28\n",
       "20. 29\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       " [1]  1  2  3  8  9 10 11 12 13 15 17 18 19 20 22 24 25 26 28 29"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trainIndex <- createDataPartition(points_data$class, p = split, list = FALSE)\n",
    "trainIndex[0:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>164</li>\n",
       "\t<li>3</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 164\n",
       "\\item 3\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 164\n",
       "2. 3\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] 164   3"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_set <- points_data[trainIndex,]\n",
    "dim(train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>86</li>\n",
       "\t<li>3</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 86\n",
       "\\item 3\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 86\n",
       "2. 3\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] 86  3"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_set <- points_data[-trainIndex,]\n",
    "dim(test_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = train_set, kernel = \"radial\", cost = 10, \n",
       "    scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  radial \n",
       "       cost:  10 \n",
       "      gamma:  0.5 \n",
       "\n",
       "Number of Support Vectors:  30\n",
       "\n",
       " ( 16 14 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_model2 <- svm(class ~. ,\n",
    "                     data = train_set, \n",
    "                     kernel = \"radial\", \n",
    "                     cost = 10, \n",
    "                     scale = FALSE)\n",
    "\n",
    "summary(svm_model2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAABaFBMVEUAAAAOHBwPHx8QISES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\nAUQCUACRABRAJAAF/j/m6Hj4LVA1pwAAAABJRU5ErkJggg==",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model2, train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predTrain <- predict(svm_model2, train_set[, -3])\n",
    "predTest <- predict(svm_model2, test_set[, -3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 80  5\n",
       "        1   2 77\n",
       "                                         \n",
       "               Accuracy : 0.9573         \n",
       "                 95% CI : (0.914, 0.9827)\n",
       "    No Information Rate : 0.5            \n",
       "    P-Value [Acc > NIR] : <2e-16         \n",
       "                                         \n",
       "                  Kappa : 0.9146         \n",
       " Mcnemar's Test P-Value : 0.4497         \n",
       "                                         \n",
       "            Sensitivity : 0.9756         \n",
       "            Specificity : 0.9390         \n",
       "         Pos Pred Value : 0.9412         \n",
       "         Neg Pred Value : 0.9747         \n",
       "             Prevalence : 0.5000         \n",
       "         Detection Rate : 0.4878         \n",
       "   Detection Prevalence : 0.5183         \n",
       "      Balanced Accuracy : 0.9573         \n",
       "                                         \n",
       "       'Positive' Class : -1             \n",
       "                                         "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTrain, train_set$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 42  2\n",
       "        1   1 41\n",
       "                                          \n",
       "               Accuracy : 0.9651          \n",
       "                 95% CI : (0.9014, 0.9927)\n",
       "    No Information Rate : 0.5             \n",
       "    P-Value [Acc > NIR] : <2e-16          \n",
       "                                          \n",
       "                  Kappa : 0.9302          \n",
       " Mcnemar's Test P-Value : 1               \n",
       "                                          \n",
       "            Sensitivity : 0.9767          \n",
       "            Specificity : 0.9535          \n",
       "         Pos Pred Value : 0.9545          \n",
       "         Neg Pred Value : 0.9762          \n",
       "             Prevalence : 0.5000          \n",
       "         Detection Rate : 0.4884          \n",
       "   Detection Prevalence : 0.5116          \n",
       "      Balanced Accuracy : 0.9651          \n",
       "                                          \n",
       "       'Positive' Class : -1              \n",
       "                                          "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTest, test_set$class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Usually, the model is expected to do a better job on predictions for the training set; because that's what it has learned. If it has generalized well enough, it should also produce good performance for the testing set. The problem still continues though; how do we know that this particular training set represents the class distribution well enough? We need to repeat this split test a number of times and compute the mean and standard deviation of the accuracy. We can do this with more random splittings, or we can use the cross-validation approach."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cross validation\n",
    "\n",
    "k-fold cross-validation splits the data set into *k* subsets (folds) and then picks one subset for\n",
    "testing and trains the model with the remaining *k-1* subsets. \n",
    "It does this for each fold; so for k=10, it'll end up doing 10 training and testing sessions. \n",
    "\n",
    "We can use cross validation to **tune the parameters** of the SVM model to get a better accuracy \n",
    "without the danger of memorization. We will use svm tuning functions for that. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "set.seed(129)\n",
    "\n",
    "# Setup for cross validation:\n",
    "# sampling=\"cross\" for cross-validation, \n",
    "# cross=10 for 10-fold,\n",
    "# retain the best model and save the performance measures\n",
    "\n",
    "tctrl <- tune.control(sampling = \"cross\",\n",
    "                         cross = 10, \n",
    "                         best.model = TRUE,\n",
    "                         performances = TRUE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Parameter tuning of 'svm':\n",
       "\n",
       "- sampling method: 10-fold cross validation \n",
       "\n",
       "- best parameters:\n",
       " gamma cost\n",
       "   0.1  100\n",
       "\n",
       "- best performance: 0.044 \n",
       "\n",
       "- Detailed performance results:\n",
       "   gamma cost error dispersion\n",
       "1  1e-06    1 0.572 0.04638007\n",
       "2  1e-05    1 0.572 0.04638007\n",
       "3  1e-04    1 0.572 0.04638007\n",
       "4  1e-03    1 0.572 0.04638007\n",
       "5  1e-02    1 0.356 0.09512565\n",
       "6  1e-01    1 0.064 0.05399588\n",
       "7  1e-06   10 0.572 0.04638007\n",
       "8  1e-05   10 0.572 0.04638007\n",
       "9  1e-04   10 0.572 0.04638007\n",
       "10 1e-03   10 0.356 0.09512565\n",
       "11 1e-02   10 0.136 0.09276014\n",
       "12 1e-01   10 0.048 0.04917090\n",
       "13 1e-06  100 0.572 0.04638007\n",
       "14 1e-05  100 0.572 0.04638007\n",
       "15 1e-04  100 0.356 0.09512565\n",
       "16 1e-03  100 0.344 0.10699948\n",
       "17 1e-02  100 0.052 0.04237400\n",
       "18 1e-01  100 0.044 0.05146736\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# now run the tune function to tune the parameters of the model \n",
    "# tune function will try to find the best parameters (gamma and cost), that means the parameters with the smallest error.\n",
    "# it will try different gamma and cost values given as arguments (e.g. cost=1, cost=10, cost=100, etc.)\n",
    "\n",
    "tuned_params_cv <- tune(svm, \n",
    "                           class ~ ., \n",
    "                           data = points_data,\n",
    "                           kernel = \"radial\",\n",
    "                           ranges=list(gamma=10^(-6:-1), cost=10^(0:2)), \n",
    "                           tunecontrol = tctrl)\n",
    "\n",
    "summary(tuned_params_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = points_data, kernel = \"radial\", cost = 100, \n",
       "    gamma = 0.1, scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  radial \n",
       "       cost:  100 \n",
       "      gamma:  0.1 \n",
       "\n",
       "Number of Support Vectors:  37\n",
       "\n",
       " ( 20 17 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# best one is with an error of 0.044. \n",
    "# cost = 100, gamma = 0.1\n",
    "\n",
    "svm_model_cv <- svm(class ~ ., \n",
    "                       data = points_data,\n",
    "                       kernel = \"radial\",\n",
    "                       cost = 100, \n",
    "                       gamma = 0.1,\n",
    "                       scale = FALSE)\n",
    "\n",
    "summary(svm_model_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAACBFBMVEUAAAALFxcPHx8SJSUU\nKioVKioZMjIaNTUbNzccOTkdPDwePDwfPj4fPz8hREQiREQmTU0oUVEqVVUsWVktW1suXFww\nYGAwYWEyZWUzZ2c0aGg2bW03bW05cnI+QkI+fHxCPDxGjIxJk5NKk5NNAABNJk1NTU1NmppQ\noqJRoqJTp6dUp6dVqqpXr69Yr69ZsrJbt7dccXFct7dfvb1jxsZjx8doaGhogIBo0NBs2dlt\nkZFx4eF16el2PDx48PB8AAB8Pnx8fHyA//+MAACMRoyMjIyMu7uNdXWPRkaaTZqampqccXGn\nAACnVKenp6euuLivtrayAACyWbKysrK08PC2d3e9AAC9X729vb296enEPDzHY8fHx8fMu7vN\nXFzN2dnQm5vQ0NDRiIjVQUHWk5PZAADZbNnZx8fZ2dnbgIDevb3fcXHh4eHjsrLnp6fpAADp\ndenp6ensmprvRETwePDwjIzw8PDyZWX0fHz6PDz7TU3/AAD/Cxf/EiX/FCr/FSr/HDn/Hjz/\nIUT/IkT/Ikb/Jk3/KFH/K1b/LVv/Llz/Ll3/MmX/M2f/NGj/Nm3/N23/Pnz/Roz/Ro3/R47/\nSZP/SpP/S5j/TZr/TZv/UKL/UaL/UqP/VKf/Var/Vqz/V6//WK//WbL/X73/Y8b/Y8f/aND/\nbNn/ceH/den/ePD/gP/////Qu2HYAAAACXBIWXMAABJ0AAASdAHeZh94AAAgAElEQVR4nO2d\ni58lx3XXCzsxCDvxOER52EmWsa1RYgdEsuygJIrxDB7AQzwQr4lHeEEsAVuIkfEIexBqQ15s\ngjEGLS+F4JBos7blvf8k/X5WdZ+qOqequu/v+7Fn9t65t8+93eerqj5VXa12AABvVOwPAMAW\ngEgAMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAAD\nEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAA\nIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxA\nJAAY2EeR7p8fKXV8cVX8+0Sp+90flDrZqYLr8vF1+W/TZmb+NPPqu3nsc9Kbzx2iaEPqNwxY\n2UORLlTN3V3pzln7l9Kq8k/3ysf3+EW6W2yRItLLR8ohCuEDNhsGrOzfTr1ULS/nD4+a5qds\ngI52lUgn5RMn/CIdK3VFenPzAnaRPDYIzOzfTs1T+eImt+akaovyJuKi/stF2Uipzh7FKJLt\nmxjyHSIFZP92apNIN9U/bspmqCRvnG7Kv59VJ073y38N9tDNRe7hyf3Bhu4VLzq+KNu1m7tF\nK3ZW9QwHD+pXjxua/gZ7W1Ktw+0HuH9edArvd1/ifr7186vxV7vMt3d+PYjRf2u3YcDK/u3S\nPNHOXu49PmvKDfeqJipPs8vyNGZ3Xv6rv4euj6o8LNuw+k8nvY5i8/eyazh4UL164sdgg70t\nTUVq/lid0RVvqR73Tar+G5BzdL0zvBUiCbF/u7Q8Rzo6v9ecGeXtznH5j+PKqEKks7KVOlJn\nI5EaNZoXVps7uSm7hWelenn7c5Nn7uXogUGk/gb7W5q88Kx5YWVS+0j1S3Dds8c7w1shkhB7\nuEubxDquu0n1+f9V3ccr/it/WbQKL+cGXA2SLn/66KpU43jX5OlxXaxoRCke3JQvGDzQd+0G\nGxxsafTC+4XeN3lnsZP46H7pav8AFs/m7eL9o77pmrfy71Owjzv1/nGtUlVluKz+cVG1HKVI\n10XZIc+866FIdS/w5vju9W6ckeWjooFpTmOGD/QiDTc4eOXwhef1hys+5Xn1bPnGyWcoT8ju\nty/Sv9Vlp4F59nOnXt87LztVZYbdtJl7s6t+X+USHOctxNFuKJJGnWprFydVf+lu3d0q9Rk8\n0Is0TupuS9MX3lSv0P118on6jaDprYCV/d2p12fNyVH53+zL5nSjFCl/qvzPOkGke037Vjxq\nBnvLs/3BA4JIgy2ZXkgUSS2+FbCydzv1qPkPdJdSV6psf9TLzbNXRQXvuOgnLYtUzH44Pm9P\npm7uVTWyk/GDZZGGWzK3SEeTv/Y+0c1k66a3Alb2bqeet4Wu6zaljstictU8VSJVs+yuRyKd\naM6R6lpfPz/vn2seaEUabHC4peELzwwnOhORylfdb0vdM28FrOzdTi2KWOWA5f2jVqlqTl2V\nbnWxrehlHfdlK9BV7eoXVMYdt+dZR6MHhKrdYEvlo5v2habS20SkotpQfLPL2bfe7AAzeydS\nb1ilm2XXnJjU/95V5zcXu3GmtsM+XZ6elK8rcldV08evm2lHgwd6kQYbHGyp/Evvfe1gbSW/\nWaSKo53xrc2GASv7J1KXWEdtbfqil1zdf8enoy4vT2c2vNxsrDzHauoLJ81W2wd6kfobHG7p\nXA3OrdpPPby4YiJSFVM/s6F6a7NhwMoeilRdj6TO7nYdnKI71cy1qdKvqIl3PauWm4v8rWeD\ngvbVeTFRohh6OtvVp0Qnl02g7oFepMEGh1s6G53SlJ+6P9eu/7t79t6xOrq4Gf5x8NZmw4CV\nfRRps4ysBwHBnt8QECke2PMbAiLFA3t+Q0CkeGDPbwiIFA/seQAYgEgAMACRAGAAIgHAAEQC\ngAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgA\nMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkA\nBiASAAxAJAAYgEgAMBBAJAXAAtIpJZLYw48UIMQTXx59K/vWI++tdGRZ/xeIjrVIwtt3YA0i\nvZkVvOm7mQ6IlBgQiRTCcy8/zh5878n3HmSPPbfT8TD7Tv7zO9lDti0CLyASKYTnXm7Snq9J\n4lcTeAGRSCE897JAR4y9swi8gEikEJ57eU4k1zLEowe85QvgBUQihfDcyzMioWXZBhCJFMJz\nL5tLAzjX2QgQiRTCcy+bdeEvQ4AoQCRSCN/dbOzAYTxoI0AkUgjv/WwqDUCkjQCRSCFE9n0B\nRNoIEIkUQmTfF2CGwkaASKQQIvu+AFW7jQCRSCFE9n0JxpG2AUQihRDZ9xWYobAJIBIphMi+\nLyYHPWS+TokYFuZyA5FIISR2fdWpC96xQ19SBIhECiGw54syw+Ps69njsKUGVDdkgEikEAJ7\nvih8P8z+OG8bgk4QwpwkGSASKYTAni/GYPP/17+CgRFgGSASKYTAnodImwIikUII7HmItCkg\nEikE+fvSa8vNOdLDsBOEMCdJBohECkH9uha1ZVTtNgVEIoUgflurLMU40paASKQQxG/bry0v\nd/KKyUEPw08QwpwkCSASKQTx2/bO5PEf/v0CIpFCEL9tJxJORfYMiEQKQfy2nUiYQLBnQCRS\nCOK37WrLGK7ZMyASKYTp640qCl2HDiLtGRCJFMLw7SYVhfYJiLRnQCRSCP2X01QUmtpy9AkE\nuHwvLBCJFEL/5WYqCrGrdqi+BwYikULov9xc/y1uJsf2eP+ASKQQ+i83eyIUdQIBqu+hgUik\nEPovx1BREDqVQa0jNBCJFEL/5fwrClIdQIgUGohUbKFi5gX6L9eeibi2K2KnMhApNBBpt7ty\nFalpUJzbFbFTmejV970DIhUinS2FMH29sqJAaFcMTZZYw4GqXWgg0m53qe4uhZj9jsvtiqnJ\nkuuBYRwpMBCpEOlyKcTsd1zUwdg+CJ7K4PK9sECk3e5M3T9XRxdzIWa/46IOxiZLeyqDyT1r\nBCIVIpWcjDbbZzaxF0UyvkDXVKFTtkogUqHMvd3u5mKmg6dmE3uxRGY2bWoNygTrBCI13Khj\nc4jZxF7M/Zkma3Iqg8k96wQiddsxb0jNJ/ZSb8xmVAdDqesEInXbmRFpIbEXSmQ23bVWJJQc\nVgVE2u2O1E3+83pmWHZJpCUsCgitSCg5rAqItNtdqIuy2HDfHMK3q0Uf1am6gX+U/RatDUOt\nPBEgUu7QUVninhlIUuFmrVXdwP+Q/Z8nlJIDauWpAJF2RWt0pI7nZjfMV+14qZcEL/+91A6i\nVp4MEIkUIsh/9OtuWtENJIqEWnkyQCRSiBCnIf1uGlEk1MqTASKRQojs+yGDbhpx5AkiJQNE\nIoUQ2fdDBt004skPREoGiEQKIbLvhwyloJXjcCFsMkAkUgiRfT9k1LqQRp5QtUsGiEQKIbLv\nhzh10zCOlAoQiRRCZN8Pceum4ULYRIBIpBAi+34IummrBiKRQrjvYM1kOMP8OHTT1gxEIoVw\n3r8aO4zCoJu2YiASKYTr7tX019CF2yQQiRTCdfdqJsNhftwmgUikEK67V1PTxmyETQKRSCFc\ndy9E2hcgEimE6+6FSPsCRCKFcN29mlFWzI/bJBCJFMJ196Jqty9AJFII5/1rM44EVgxEIoVw\n38GaUVYMvG4QiEQKIbLvwYaASKQQIvveASxjlyoQiRRCZN/bg9OrZIFIpBAi+94aFPzSBSKR\nQojse2swTS9dIBIphMi+twaTItIFIpFCiOx7ayBSukAkUgiRfW8NREoXiEQKIbLvDTz8rey3\n9BPxME0vXSASKYTIvtfzO2WJ+3d0f0LVLl0gEimEyL7X8ofZ1//syZ99PftD3R8xjpQsEIkU\nQmTfa3mQ/XH+84/1TRKm6SULRCKFENn3WqpKwpvlr+9mb4ULDLyASKQQIvteSyXSO+WvBw++\nHy4w8AIikUKI7HstbYn77Sdvo6qwHiASKYTIvtfSnCP9++wRytwrAiKRQojsey1N1e6/Z1n2\nTriwwBOIRAohsu/1NONI386+HTAq8AQikUKI7HsDD79ezGx4By3SqoBIpBAi+36Wh/k5EkZe\n1wNEIoUQ2fdzPM47dm9nGHtdDRCJFEJk38/w/QcPnhQVvACdu0+qL+Y/v6g+LR9qy0AkUgiR\nfT/DW9l3nxQzGwJ07r6insl/PvPMV+VDbRmIRAohsu8T4fPqs08+q74Q+2OsHIhECiGy71Ph\nk+rz6lOxP8TagUikECL7PhW+opT6SuwPsXYgEinEqTwih5fGZ9RnIkbfBhCJFCKASPFMQovE\nAEQihQghUjSTPpWfI30yVvCtAJFIIYKIJGXSwoLhX8g7dp9VnxcKvi9AJFKIMCLJmLSw0MNX\nnynHkbw6d1Jr+6/ongEQiRQikEgSJi0tPfTpemaDR+dOak2WNa31ApFIIUKJJGCS/ILhUquE\nrWr1MYhEChFMJC+TtNPm5JdnlVJ1VfcMgEikEOFE8jFJO21OXiSpCKtaoRkikUIEFMnHJN20\nOYgUBIhEChFSJB+TNNPm5BcMh0hPIBIxRFCRPEzSTFKQP2WXUnVV9wyASKQQYUXyMEkzba4q\nIn9DbkgGVbsnEIkYIrBIi5h2t3baXLFg+Dckh2QwjgSRiCFiizPBsLtN0+aE/+Mutbb/iu4Z\nAJFIIWJ7M0W7t43T5lY1JLNKIBIpRGxtNGh2tnnaXFH5ejP7bv4Ld7iQASKRQsS2Rsd0Z5un\nzRUivZM9yH/hDhcyQCRSiNjSaLE4DOVYzKP8zB13uBACIpFCxHZGD/0wVOdI38z+yzoGZVYI\nRCKFiK2MAfJhqKp2/xHriYsBkUghYhtjgnwcqiGZ38cdLqSASKQQsYUxQj4QxZDM/xNrkVZ0\nKasQEIkUIrYvZmwOhtgdLtY0BUEIiEQKEVuXGejHQuwOF6uaFCcERCKFiG2LPdND4XuHC3P3\nDfMmIBIxRGwtHJgcCs87XMx031Z14ZAQEIkUIrYVLvAmylz3jVukNZYuIBIpRGwpnGBNlLnu\nG7NIqyxdQCRSiNhOuMGZKHOy8F7Kus7SBUQihYithCOMiTInEm/qr7N0AZFIIWIb4YptOrxZ\nFySml1rMdt9YO2PrLF1AJFKI2EK4YpsO72RliVxzqcV8923hUlar6gFE4ti+AxBpBtt8eJS9\n/UR7qYVP982uwYJIHNt3ACLNYZsQb2aP9K2Oe/fN0sFVrcLVApFIIWLr4IHlAXvHOLHVeSUS\ny+oBqnYc23cAIs1jecS+zX6phW1fDeNIDNt3ACItYHXAzC2SM9YnPStahasFIpFCxHaBi+UD\nJnCpRbVgxLcyiYnnyQCRSCFiC8DG0vHyv9RiWuouzpFW2V2zASKRQsTOfz7mD5fvpRbaE5zH\n2YM/yX7397I/XVsBwQaIRAoRO/0ZmT1cnpdaGEpujVxrm/ZjA0QihYid/ZyI5FGDvtT9KKu6\neysbZLUBIpFCxE5+VkQSqcZQoVvnbAUbIBIpROzc50Ukkyq0xrQVO4jUpZTw9h2ASLaIZFKF\nTqS2/rC2aT82QCRSiNipz4xIKpVoJsoV9YfH2dezx6ub9mMDRCKFiJ353IjkUoGmale6hXGk\ncUoJb98BiOSASDIVTI2punlN4W6rQCRSiNh5z49INhVMJsptv2JXAJFIIWKnvRu31C/lP39J\nPav7o0g66YBI2pQS3r4DEMnEC+og/3lw8ImF14lkVss6L9SzBSKRQoRIewFuq+dOn1N3Fl8n\nkloN67xQzxaIRAoRIOlFuKVuq0PC60Ryq2HzFbuCqCKpIa5Z7vg+mxDiGS/EC/lufYHyQpHk\naljjhXq2iIs0Z8olRJLm4+rjtBeKZNceEbdrd3V04p7d7Ufy30TOy3ObWa1I5BYJJnkS+Rzp\nSl1YfgDNR/LeQs7N0SZFOszPkW4RXyuSX3tD7GLDpbqy/ATTj+S7gYKz2Z7lWkW6k3fsnlO3\nF0aUGkQSbF+ILRIDHCHuzZ+irVSkTxyU40h55442oiSSYXsCRCq4VidbE+n17KXTZ9Xfz17L\n26Fb1BElkRTbDyBSwYm63ppIL2Zv5D/feONz9WPaiJJIju0FECnnrrq3m4g0qMyL5rwMr2Sv\nnr6afal5SKzfiSTZPgCRitLh2W4q0iCEZMZL8Xr2Sva17iF5RKlFIN3WeHdYGhBptzs+utmi\nSC9mWfZi+8hiRKmFPds2PFkIIu3O1f3dFkU6/XL25e6BzYhSC3OybXn6KkTaEeYorVKkQYvU\njijZ0R5H8z0xLVjn3WFpQKTNivS1/Bzp9frf3YiSJc1xNN8T04ItX+IHkZqtbK1r96W8Y/dq\n9kr14Nl6ZkPVuSPNc6hpDqTxnpgWQKReSglv3wGIpONzb5TjSL1yQwf1ytmS5kga74lJByL1\nUkp4+w5AJB2vZS/lP1+qO3ev149eq/5IvXK2oDmSbncgG9S7t3zROUQihZBJ9nC4zXOoaA6l\nyz0xh/VuVO16KSW8fQcgEgG3eQ4V1ZF0aZHG5mAcqUsp4e07AJEo+MxzKI+kyz0xJ/Xu7V50\nDpFIIfgTOzRe8xyeON4Tc8vVhREQiRRCILND4zXPwfGemBDJnFLC23cAIlHwnOfQFQFtjj5E\nMqeU8PYdgEgUqnkOtz+oPnjbfZ5DicXR33K9ewREIoXgTeoIVPMc/kk5E+rWcJ6DNfSjv+V6\n9wiIRArBmtR8VA0MgWqew7/L/u6vnP7KAXEk1gz98G+43j0CIpFCeGaeELeqBsb8gla06hTn\nH2b/Nv/5vGtL1EE//tutd4+ASKQQvpknwh11MN/AjEWrL5lnuHJ++cBv91pYPRCJFMI78yQ4\nVM+fzjUwE9H4RFo0aX/6dDUQiRTCP/MEWPJiIhqjSAsmUasM22m3IBIpBEPm8bPkxeTvjVnU\n6aqzzB524rWwG2q3IBIpBEfmsWMt0uJJlRVzh502Erul6jhEIoVgyTxulhqYqWiLZT4azdVN\n5gUcaCJtaQ0HiEQK4Zt6Iiw1MJ1obRn89gFx4KnH6KLA0+7qJvMCDjSRtjSDCCKRQrjkuTwL\nDUwr2uR1g4HcpRUcxhcFnrZXN5kPO21uEESS274D+yvSUgNTCzRpuYZiLa7gMLoosKC+usl4\n2GlnPxBJbvsO7LFIS1SijcvgY7EWV3AYXRR42s0lNx53Uj1uS3NaIRIphECSh2NcdJiMLy2t\n4DC8KLCkubrJeOApc4NQtZPbvgMQaYGxSJNq3uL1soOLAgs6tbyyD+NIYtt3ACItsCiSYQWH\ntgoxbZF6q7h6pd925rRCJFII5tQOy3i8idoitVWI/uLHJYNVXEXycnVAJFIIifwOxri40Ij1\nb5oBItMKDnUVYqBNwXAVV5G8XB0QiRRCNtOlGY0jNWL9g3qA6G8bV3AoqxDTxY+Hq7jCpAKI\nRAohmufyjMabGrGqAaJ/bF7BoezzjbTRIJKZKwMikUKwp7YYpKvPG7HKAaK5FRyI60iKpOa6\ngEikEJbZHI+qsXkPcUadZoBoAHkdSZHcXBUQiRSClMQJUJ7+fIg+x3syQDSEvo6kSHKuCYhE\nCkFKpgQoCnJ31PvULdpVRwstksU6kiLJuSYgEikENZFjU4wN5TLlv0hrBU0GiAZYrSO5mAks\nd6FNF4hECkFN5NgUIuX/r38tMRkgGmK3juRSJrDchTZdIBIpBDmTI2Ml0tzdMR1YSgWOu9CG\nwn5RFohECsGTavIU50hKlbOBlkVaHiCyYykXGO5CGwiHybQQiRSCKdXEKap2h+qH1B2utYKs\nWMgFt7vQRsDl8g6IRAoRPCddqcaRfsq8lMNgwJa8djiRhWRwuQttDFwWZYFIpBCMySbM7QP1\nnplxpP60u9vVC5/ms2k+F1bTIrlcAg+RSCF48iwU5qUc+hPB68br3TxLdFXM5oLLXWijAJHE\nQvCkWQyGvbdD9Wz++NnCmzvqB9XzuVOzHUFrZlLB6S60UYBIYiF4siwCoysoVEPhVPHzeVX/\nYmqSzCa53YU2Ci6LskAkUgimLAvO+Jo+VT9WlVOnvV9cIU2Z8FY9s2EFnTtU7cRCcGVZaMYL\nBqn6saBIW5h2h3EkqRBsWRaY6bonPZFKy3KPDk8h0gj7RVkgEikEW5YFZipS27U7VD+bPyjO\nke6w3eqlwjJFtgFEIoXgy7KwTEVqiw356VN14dK7GW/1UiGSqYkDkUghGLNMlPFUhfFKXIfq\n2YOi/F08ri+mfVo3juQ55UEkVdMGIpFCuOdUUCa3nWiqdh+ovRhU8epxW83wrfd9lERyNWkg\nEimEe0qFRHPDpFttZ67ygqQIw539RJI1ZSASKYR7RoVEd5vz20VzNG2GrLdji20irv22zBCJ\nFMIjowJiuKmstRcsdz+3y5PVL6cPkUghfDIqHAYBrL1gEcnKpPXf4AUikUJ4ZVQw0hLJxqT1\n35YZIpFC+GVUKAy3Obf2Yul26VToabL+m2BCJFIIz4wKhKHaZu0FQ9Wugpwm0iLJlzIgEimE\nb0YFQl/ctvfCexypgZomwiIFKGVAJFII/5SSpZyJkP84eEpX3DZ7YZrBQCmSU6CmiextmUOU\nMiASKQRDTkkyHnYdY/KCreUxQU0T2VQPUcqASKQQcqnGQdl3+ytK/VXLExu2cyEj5DwR7XyF\nKGVAJFIIsUxjoawmHJZNi3nYVbMOF8cMhnnoiSJ5W2aIRPtIAUKIZRoLlEtdB724+gHTeNEM\nIjlrDUSifaQAIcQyjYVWpLyhMVjR9OI+UjRFzYOeSNxLRdaI5Kw1sqWMCohECsGfYpy0Iplr\nB00vruT97YPm7dOqA49ZIjlrDap2tI8UIIR3RonSnCOVl47/gLqjkaBy5o5SzZXm5XO1UD/5\nAfWujw2rDkz1PJGctQfjSKSPFCCEb0KJ0PrSVO3KBR9/Tr1XI0HlzmH56/lOpKaLV76+V3Xg\nqueJ5KwDkqWMCohECuGZTyL0Go3BOJLSSVC5o5pf9XP1+3KVytd351f29bxb9V3Jnh08K5Kz\nSQKRSCGcMl2WqtF4SnWXi6vq36onQdtoVWZUS2+1rzgs3/fB9laZnUj29bwXVHmfzINPDJ4V\nydkkgUikEH45L0KpxuBUZpj9xa/uz5V2qmqmhm1Wd4c/H5FOb6vnTp8b9wVFcjZJIBIphFOq\ny1IvqVX3ygoOu9Ufq79P7z1Rv+L9I/96N/rrNn5qWRi/pW5PJpiL5GySQCRSCK+Ul2F6A/M7\n3eqPp60f7ZlO0Yv7cGNWf/JdbVxxqtS2J92VF/Ty3Qv56yY3QBdJ2hSBSKQQfjkvgua+y7cG\no0SHmg6aTotCmvHzbVtmU777uPr49EmRrE0QiEQKYZvlAdDdd7lsaLrc15zpaOaBl6//2J9T\nH+g/35hlUb7Ttkh7YxJEIoWwyfBA9Lpjo5pA27wQSwb16dMPDhSrjbOoOhzm50g630TyNjkg\nEinEchqFp/VlchF50+xQrzG/fWCcXkQX6U7esXtO6coSIombGhCJFGIxjWJQ5P9Tc2cw5BOc\nO+q96uCvH6i/OHkpWaRPHJTjSLrO3V6YBJFIIZbSKBYLNTVqye1Q/XTeeD2vfmLyWvLCKc/W\nMxv2tXMHkUghFlMxFuPiwWjUh7j2Qnst06Tl4Zl0J5K6zogsKQSRSCG80iggxibIOKxa/sEs\nEtM0cO609UFmKjhEIoXwzCNWZqYaGNsPow7ddNdTvUhMCwoxp60HhouTfJspiEQK4Z1IfMw1\nEdNRn8q6wQTXPu1kiOoc6adnz4W8rvXzyFFe9EsKeTdTEIkUwjV9+Jk9aZnU2GrrphNca9rp\neaaqXQ/PTp57jvKiXcDB/xpaiEQK4Zo9/MxONagnmbYi9RdnGE5w7b3hdOky9eG2XMsOzinK\njFYk/5XvIBIphGPyCKCb9tN2ucarLxz2FmcYTnAdby0/E3rXfMdtaHAZ8/XspfyfL2WvUT64\nc4oyoxXJf50hiEQKQcmUMIxEyhP6PZ06h+rnynUbfrC71Kj6NZ3gqt0aOXCl69/L3sj//cYb\nn6N8cOcUtWOxaKBdUggi7fZbpCqhf6rpct1RP1Cu2/CzdbvRX+VkPMG1xOJGFf3ATTfvn2ev\nnr6afYn0wZ1T1IrlooH2dAgi7fZNpEHqFwl9qH4ol+j54b2WxxfLmia4Wpz49DfZdvNez17J\nvkb74M4pagOlaKBzzX/lO4hECkHLlRAMUr+5lKI/07s/abu1zjjBlV6K6xvcBnsxT8kXaR/c\nOUVtIBUNNEsKFQI++k9Z9r+cI0MkUghargShn/rTc5/hr8464wRXzXirfryob3AX5cvZl4mf\n2zlFbTB30SbnTsMnqmbKYyQJIpFCEJMlCL3Un4o0OunpW0dtfEyv6z2fbItkFGnSnxs/8T+z\n7BuPPEaSIBIpBC1XgtMva1fqjE96+g0ObbKP+bSpe3+r69fyc6TXaZ/VMUHtMIk0OXeaPOE7\nkgSRSCFouRKcn1TFgibvU4dd4nuv4k25uLyR7Z/mHbtXs1dI23VMUDtMRYOJJsUTeecu+0bz\nhG/dDiKRQtDzMCSVNH++UacyZtzuzPfoppaRhpaqjf6dN8pxJFrnzjFB7TBV7Saa5P8adu4g\n0v6KlLcLHyoT+t2lDC53Yta8hzZGW+r6Wj2zgdS5c0xQSwzjSBqRSufy1z42vMISiEQKQcmU\n4BSdsKIa93T5yGCMsadWtEUf0bxH6u5jjglqi365fI1IZW+vk853JAkikUIwJxUPw5Q3GGPy\nor4Q6Uduj99jMdnBCscE5WGiycOskCp/onHLd/43RCKFYE4qHrRDR2NjDE8X7dd42a7mL089\nrZ5+iv0OzY4JyoOmapcbVDzRNlKeVyRBJFII5qTiwUekvNAfy0AAACAASURBVN0pr6v40PB2\nLgXV7ZXey/1pXTOUB+M4Utfb87uHEkQiheDOKifGFbZhJ8xszKmmp1atDK6q+UXjFumH1Y9u\nrEXSaPKtrHiC7cayEIkUgjmpnJhU2JrywgdKvwzGGGoQ1XyIw/LX8D329xij4ZqhImv+PFk4\nKbIPCpFIIZiTygWNEYMb9Znq3PqqeCVScU/ZwV0oTpOr2rmdulBEmNmyQ1CIRArBnFQu6FqK\n20Vz1PhjGnnVzgxqJhbpxpH688f5sEycGrdiGk0E40mRS1CIRArBnFQuLJwD3SqNOehPrZub\nGpS3X7fULx6oD79PPXW7/+L+FU1eywaNsUycGqcpcL6lbJegEIkUgiubPKBU5Swme9d/HtzP\n71b/Glube4xRsEycGqcJB74TUF2CQqScm3Olzq/mQjAlkw8EkfqnSYvXvja3oLg1eHGz6sO7\n7e4xRsAycWqcRNK+yaJ+AJEcOSozasakFEQy17HbX/3TKErxrT176l7cW4eIuXxnmTg1bCLZ\n1A8gkhsX6rz4cTYTgieXvFiYSze4AvzUsvjWvbg8NTpolh2ib2EZy8SpcZoCpxHB6rTJJShE\nKhqkm2IzM9tJQSTC7G4mkU7bS2/pW1jGMnFqLATo+m4aEaxOm1C189rM0czfeHLJE/0Vrror\nwE+dReo1cE4i3apvk/Ts5C+WidNA7pL1XqgRwa63hnEkdy7U5UwIm1wKjeYK8FPLSdzdi3sN\nnNM08BdUeeO+g09M/mKZOC3EKXADeaYiEEVqWjX7eXcQqeSeUhfjzfaxyaV4WFXtDG+8pSnl\n2XBbPXf6nO49loljy7DvNhGBJpLHDHCIVHJ5dqTuzoSwyqV4uCwaVI669l7cNXBu40i31G1t\nI+aSnRYsmEKqH/gM5EKkhvOZvt1aRHJYNKi2Rftip3uMvZBvLsIdmRdEIjniM5ALkRpuZqoN\nqxHJGuZR14KPq4/rnnZIThuW+m6UXpvPug0QqduOeUPbFYn/oolILdJi341QP4BIflTjSNfq\n2ByCLc1SQ6n6bsxsE1QP83MknZYOyWmD/z33IJLvBsqZDTdn6zxH8hRgUJvkaJfu5B2755Tm\nEzkkpxUMNyv3WUkIIu2auXYnMyH8E0wI3xnaqrk5JuVUiSDtJw7KcSRN584hOe3wW3OhAFU7\nXy6O1PHMeGy6IlnVCnQeHKifz3/+vKKcKlGkfbae2TB9lUNyBgfjSNIhKHkaA5tagdaDtkWq\nH82837PA55KdwXFv1SASKYRb7sjTF6Df4GgaH70Hvbkb/duh6/As8Dml53qASKQQbrkjT0+k\npVkNeg/K+5kXVbvljpvnbHCR9E0HiEQK4ZY78nTJvTjPTu9Bo9fodujzsZwQSd8a4tWvUit7\nFRuGSKQQbrkjTzdDe/HqWL0HjXOj26E39LuI6YpErBEwFMj1KhYbhkikEG65I0/lwVNKfXDx\nWiSDB3WPbnQ79OEfqweeq+v75O88xKo1w5CtXsVywxCJFMItd8ToGoreGpHlY7NIYw+abdQz\nU3XL2Q27iMlW7YhzTX3XFjKqWG4YIpFCuOWOFP2GorlduaoS3CzSyINxbUH3nlEX0W/w1zl9\nFyHO7PG9lZhRxXKLEIkUwjF5ZBgpUaX7T6ufOJ2/Orby4GntNvTvGcvldFlFg3P6LiIvUn1q\nZNgCRKKHcM0eEUYNRZXnxTLeC1fHPt22KMWQ0bOng3qE7j2m8oLT/D779KUiLtLoFjAQyT2E\nbdaIMsrv9uHC1bF3NKuE9xyh307WrYtnnb5kiHNNnaektqdGBpFwjkQPYZk0sphEWrg6tm3I\nCqPqe1D0HZm+R99FdCw62KYvHemqXXtqZFARVTt6CLucEWYkErUs3b6teEN1hzHDoFDTddMr\n4zhVyDZ9LfAdR1oYqG0bIpOKmx5HOmcNYZczwozMobYQvS5gd1sXnX1d180wz7X/i4xl4lhB\nnGtqeNmSh12PzvTKfMMbEqmvzstHrIHTEmlsDvGcZSBSeQPZj71LfWBaNOhvXtdFTFAkLxa7\nfL1TI7OxGxJJHb3cPHU+e5meQwi7nJFmbA6tLN02ZKUC7UjuRMClrtvWRFocqCVVKTYk0olS\nVaOUN0dzy6a6hLDLGXGcBnTalqYy5Vml3vUxXZdwSRTHqUKWiROOxbI4qUqxIZF2l0dlo1Q0\nR9fMIexyJlGahqydqKpveZZESq5q58ny+BKlmLElkYoFTNQ5e3O024pIp7efUurgdtetO62G\nZkdt22LXLbVxJE8IA7WEYsamRNrt7uZHeOY+R84hLJMmNoapB60BzcV8s1cAmrtuTj1Ly8QJ\nh8/aQR3WItnCn9WTj9T+6/qkbJGO7rGHsM2auBiajFGfrL4asB6aNb6MC56sF4BjRTwHkSx3\nX9BzJNWcI53dMIfgzSlhTCY0Tc3TVXuVP8z/l7c849Mk3pswNzClvR2ka2IZLvjblEh7VLWb\nx1S/Hk3KK4eSSt/Gp0PzXTfzvcRm8ctTN4iK+K+ItymR9mccaQFTtWA8TbxteeyGhMz3EpvF\nM1Fd4Om00diQSHszs2GJVqRRzaF6vhqPLdur29VVsbZjq8Z7ic0SIJnH+F8TS2dDIg3Z8Fy7\nJQZduF4Hr+ryqaocV77oUP1IrpL12KrpXmKzBEjmMf7XxNLZrEi8IWyzJhTaOne7xNao5lAV\nIfrXpP94qdq7bVsX051bZgmQzGMgkt1HChDCNmsC0ZzlDH0yzFwoV1Ktx2Erke7kDhX8mG1c\nw73EZgmQzGMgkt1HChDCOm2C0BjzY6M+3K2xMb0nP/iT3Xhr3nTdPsh7d7aVbpcWKUAuT+AZ\naqUBkUghLL91IEx9uG7mwmkrUju41Btlcl3y0XQvsRkCpPIUVO3sPlKAEJbfOhCtCNpxo6Em\n3eDSZMUGW5GM9xIzEyCTdbAMtdKASKQQlt86EEMRxj4M58z1XtSOt7qJZL6XmJEQmayFY6iV\nBkQihbD81oGYF2k4U2huAUi7Urb5XmImgmTyALHl8o1AJFIIy28diPYcqXw0aVgGc+aGL6rK\nfK1qbHdi1hI0oUsCdukaIBIphEiCeTOqc08alv6cuUHr0yg2WkF/EadpdiHTuSJkkaEBIpFC\nWH7rUFQG/Djlugf9/ZNK1ejXTbhMswuYzA0hpwY1QCRSCMtvHYyqzSG1KL0XjaaHWyxUZz/N\nLmAut4QciG2ASKQQlt86NMuXrOZnQQdPNS8yLdVKiGQ7zS5gKndAJCcg0gK5RO8ZtFmjwoPN\nHHDLSQ0BM7kHRHICIs1T9el+aniReffL8qoku2l2ATO5B+PUIHIdHSKRQlh+65QoSgmH6ody\niSYnRYfVn2+pX8wdow0n2bVI/onsBF/Vjl5Hh0ikEJbfOiU0S+b3y3T5n/OHf0196EB9hDCa\nZDXNzjuPXeEaR7IwEiKRQlh+65Qo7Mn/X/+q6FXwukW53kuo/VlNs/POY3eYpgZZ1NEhEimE\n5bdOCZ1IvTJftUzkgVIfrpupuXbJapqdfx5Hx6JqAZFIISy/tRfM83UWbuIyuGtSOd9opl2y\nmWYnnONBgEjcISy/tQ/cq8oVgtxR71OH+rkLw3tUlGsM3f5RpT7iG1Y4xcMAkbhD+KYVHfZ1\nTssNfshoZxuvXWOIxWThDA+ERR0dIpFC+GWVDY43lpzhVr3CiaG/2HhTiVT8OzdLKT+TxVM8\nDKjacYfwSSrLHeZ4+fcMCxOI6j83i3UdFL89myT5FA8ExpGYQ3gkle0O4xeJRrNYVxm5XTrF\nCfH8Dge5jg6RSCHck8qWJZHkrsFrb4bZqz04IZ3cSQKRSCGck8qahcu/Ze4UUVH08T5ciNTV\nHlyIndJxgEikEK5JZY+5ale0RR+RuXdRj9pU62WMO2KndBwgEimEa1I5YGpz6p4Xd01vQt4m\nPX3bw9XYGR0JiEQK4ZhUTuiLbNOFu6Xw6z3GTuhYQCRSCMesYqSpTrczTeVwukVsTex8jgZE\nIoVwTCtG2vHS7kGCxE7neEAkUgiRrLPbkc0MntOypie7Ep0rsbM5IhCJFEIk7ex2ZDunVHf/\niZa4gsXO5ohAJFIIkbSzYnCVg3ElO8lhpmViJ3NMIBIphEjeWdGML324qASYZrayTx23InYu\nRwUikUKIJJ4d/cbGNI+IZeq407rEp3vuEUSihbD81iL0qtImkVhmvLqsS3y6Fx7Nrc0FkUgh\nLL+1NDIiNYUK+3WJT/fCo9lrKiASKYTlt5bGNLPVJBKpltf1HW3XJT7dC4/mr/KDSKQQlt9a\nGlNRwSBYpch75mXqbdPhZsuBkzoG82tzQSRSCMtvLY6hzK0XbGHVhpp+ocKwLvHr2Uv5z5ey\n1yZ/CZvScZhfCQUikUJYfmt5DPPhtII16wjdmi2M97qFphbpxeyN/Ocbb3xu/IeQ+RwNiMQQ\nwvJbx0MnWH9lO3OT1BPJuC7xK9mrp69mXxo/HTKd4wGRGEJYfuu00K61qn1V9WtmXeLXs1ey\nr42fDJnNEZlfmwsikUJYfuvIjIp0uRvFPZAWRGoLFXPrEr+YZdmLo+fCpnM8ULVjCGH5reMy\nPlE6VD9RPvP+WZHaQsXsusRfzr48eiZwOkcE40j+ISy/dVRaI5qG6Y5SP3xL/QW1cA8kyoTX\naYsUNpfjMrc2F0QihbD81iGZDLY2fbTuNi3vL//1l4p7IH1gZjCJcGXs1/JzpNf7TwRN5JSB\nSKQQlt86INN2pK0atGNKxfKp71Etg0bH6gKmL+Udu1ezV7onYqdvOkAkUgjLbx0OzRBsK1I7\nwNreA2n64v68oMVJ3597oxxH6jp3sbM3ISASKYTltw6H5sKJ4YS7Xq1Os5jXYF7Q3KTvUrPX\n/kWh2Utt5y528qYERCKFsPzW4dBMU22vpW3/0l1dO35x38PZSd9azWInb0pAJFIIy28dDo1I\nTSvT/aV9ZvLiwTOzk741msXO3aSASKQQlt86HLoLJ6rznvf3ZoLXZ0ILIs1P+p5oFjt10wIi\nkUJYfutwaC+cKOvYgzJEVdmevniolmHSd8VYs9iZO3/BanAgEimE5bcOx8xyJ9PC+PTFA7UW\nLkMaahY7cW1uAhYCiEQKYfmtAzIzH2E6wDp58UAt46TvkqFmsfPW6raUIYBIpBCW3zokVit1\nT17cU2tm0nfBULPYebt0wWpwIBIphOW3XhGtWnOTvk/HmsVO2ydLlwcFByKRQjjn6XqYnfQ9\n0ix21hZAJG4gUgAGmsVO2hKIxA1E4sW8xEmNe7ZxFqznL1gNDkQihfBLzXXxC+o385+/qX7Z\n8HddGr2ZfTf/+d3srdlkYy1Yo2rHDUTyZdgGvaA++urpqx/9y/1pdUumvJM9yH8+ePD9uVxj\nTn2MIzEDkXwZLbN1W/36K78+mla3ZMqj7O0nby8owl2wnrtgNTgQiRSCJ2GTZbTM1i+o31B/\no//3J4Upf5D979wUU6P0ZvZo6XQlsfIALxCJFIIlXRNmuMyWdlpd0ZN6aG6U3sn//M58rjGK\nlNQ0uxKIRArBkq0JM1zU5MVfU7/WW+KkypTClD+a6b59O/v2Qq7xiZTW6VEJRCKF4EnXhBks\ns/Uv8xbpX7WPmlT5dvb7bfdtWnsgtEhsBevECnYlEIkUgidb02XQIn3pb6nb6m82S5w0mfJO\n2QrUskxrDw9zyRaaCLb8T2yaXQlEIoXgS9k0KZbZ+tV68sIv/6NyWt1v1GI1mVKY8vtt921c\npXuc/+XtbOG0hatHlmLVAiKRQnAnbmKUy2z963pZhn/20Z/Jf/9M3blrEqUw5f/2um/DKt33\nH5Qt1FLnjqlgDZFEgEi+1Mts/Ua1LIN2Wl1pysNeezI8J3qrPmcK09mCSCJAJF9eq2c2/Opk\n9ZM2TwpTHmf/I/uDtvs2rdLVNWn50nRi0+xKIBIpBFvOJs3kWvNBqgy7b9MqXX0GFKA0jaqd\nCBCJjdHqJ1WKNKXu/zzovk2qdE12h0hyjCNJAJG4GLVIdYpop9lNq3RNfytIaTqpaXYlEKng\n8lgdXdzMhGDN12QxLMugmZCqqdINKwBJFQJCAJFyLsoVQI7MJu2HSMZlGaYTUjVVOohkxRZF\nulLnuUOX6twcgjtnU8S8LANh+g9Esky6LYp0Vm1BmTe0FyLNLMuwPCF1co6UUmk6BBCp286e\ni9RnmCSkFilk1S5BIFLDjToxhxDJ1nQZJQlhQuqTkONIKQKRGi7V/eFm+4ika7KMcoQ0IfVJ\nW5OeL02nd0keDxCp5vrobCaESL6myihFiBNSaWy2vYJIFTdH5o7dnok0ThHOCakpTu7hASJV\nnBzPhhDJ2DQRTTeRS/KS6C1CpILr45Pr2RAiKZskduljm8MSo0xp9BYhUs79mYJdFUIkZ5PB\nOXusc1hApER6ixApb4+WPIJIeuxzWECkRBZwgEi73XlT5DaHEMnfVHBOHvscFrgkL5E5SRCp\nN2JkfoVIAieCe/LY57BAPwwicYHrkbzwSB6HHOavDEAkLiCSDz7J45LD7JfkJbKAA0QihRDJ\n4RTwSp4kchhVOy4gkgdeySOfw5RxKowjMQGR3PHMHukcpm0/iQUcIBIphEgWx8c7fWRzOJFe\nGwmIRAohksbRiZ18SyQy1koCIpFCiORxbOJknMX0vEQq2yQgEimESCJHJk7C2ZxWQaQOiBSf\n2DnWx+q0ByJ1QKT4xM6xPlanPUmMUxGBSKQQIgkeiNgpNsCqkUHVrgMixcY/Rz6pvpj//KL6\ndPOExyWpdr21NMZaSUAkUgiRFA8CQ458RT2T/3zmma/Wj33S2/K0J4mxVhIQiRRCJMdDwJIk\nn1efffJZ9YX6UdfhcmiZ1nTaYwVEIoUQSfIAmA98c9ejtyhZ8kn1efWp5kFbL3BpmdZ02mMF\nRCKFEMnyAJgPvPauRya+opT6SvOg7Z05ObGi0x4rIBIphEiWyzN35DV3PTLzGfWZ9t+dSE4T\neNZz2mMFRCKFEElzceYP/fSuR0YMLVL/4b4DkUghRPJcmoVDT7rHRMWn8nOkTzYP2vu3QKQe\nEIkUQiTRhVk89pS7HpV8Ie/YfVZ9vn7U3r+lbZm22FWzBSKRQohkuiyLh57cIn31mXIcqe3c\n1fWCqmX65kaLB7ZAJFIIkVTnwfnQ0+56lPPpemZD27mr6gVly/SnWfYnmyxn2wKRSCFEFODB\n9chT73rUMhl+rVqmbz5Zy7V3okAkUggRBVhwPfDWdz3SDAA9eoCKQwNEIoUQcYAD5wNve9cj\nw5QEiFQDkUghRCRgIFyiGC4kgkg1EIkUQsQCfwImisGYzU5CtQUikUKIaOBNyEQxiLTZSai2\nQCRSCBEPfAmaKKY+3FYnodoCkUghRETwJGyiGPtwG52EagtEIoUQMcGPwImCPtw8EIkUQkQF\nL4JnCvpws0AkUggRF3yIkCrB+nAea6vEAyKRQojIQCJ2goRnnU0fRCKFEHGEROwECc5KT8Yg\nEimEiCMUYudHeNZ0C4oeEIkUQkQSArHTIwIrnXQEkUghRCxZJnZ2xAAiaYFIHpCO4CprXDNA\nJC0QyR3SAVxnjWuGlU6DhUikECKiLEE5fiutcc2w0m8EkUghRERZgHT8VlrjmmOdbSxEIoUQ\nMWUe2vFb6RnFLKucBguRSCFEVJmFePy2KNIqgUikECKuzEE9fhApESASKYSILDOQj59UjWtr\nRXVxIBIpREA17BCqca3zhD8mEIkUIlmRZFJ+pSXomEAkUohkPZKpcW2wqC4NRCKFCO5R3HMU\nlDCsgUikEKE9inyOApGsiSqSGuKa5Y7vswkR2KPY5ygQyZqoIl3uh0j2RyX2OcpKJ47GJG7X\n7uroxD2724/kv4nFEGE9it4ixG4RV0jkc6QrdeGc3e1H8t7CcoiwHkUXKfY52gqJXWy4VFeu\n2d1+JN8NEEKE9Si+SOucOBoTcZEYToKWPpLMZgchwnqEc5T1EbtFYiBlkRyPCs5RVgdEIoUI\nfVhwjrI20hDJq9e3RZFwjrI2IBIpROzDBFIHIpFCxD5MIHUgEilE7MMEUgcikULEPkwgdSAS\nKUTswwRSJw2RvIBIID4QiRQi9mECqQORSCFiHyZnsBpQICASKUTsw+QKZkiEAiKRQrjt3OjN\nAebsBQMikUI47dv4zUHsK233CIhECuGyaxNoDuJf17Q3QCRSCJddm0BzAJGCAZFIIVx2bQJZ\nnMBH2BcgEimEy64lZbFsPQJX2gYDIpFCuOxaikjC9YgETtP2BYhECuGyawnNgXiixy8c7gsQ\niRTCZdcSLJGvR+BK20BAJFIIp3273BygGLAZIBIphNvOXWwO7EWKPlkC6IFIpBBCe99aJJzz\npApEIoVw3b0LDchsPULzXlThkgUikUI47t2lBmRODN17E5gsAfRAJFIIt5273ICYTdO+F8WJ\nZIFIpBBuO5fQgPTqEcOunPa9EClZIBIphNvOtcr7UeOkfS9EShaIRArhtnNt8n7cldO+d1Cc\nQCU8JSASKYTbzrURadyV0763bxsq4UkBkUgh3HauzeTrLCsbmFYc/Xs7e1AJTwuIRArhtnNt\nkj2rFGlFMry3LU6gEp4WEIkUwnHv0rtfj7Os9Kbryi28F3WHtIBIpBCuu5c8+fphVrQw38l6\nXbn591JEQjkiHBCJFEL8OOSNT9GVy8jnPASRUI4ICEQihRA/DuXpUZbRe2rLhQyUI0ICkUgh\nxI9DblDRlaOLtKwJyhEhgUikEE771uYUxX6ZksWOG8oRIYFIpBAuu9bqFMWhH7ZUyIBIIYFI\npBAOe9ZSDf7KAEQKCUQihXDYs7anKOzLlGBRu5BAJFKI/i4jnvpEbxBQtQsJRCKF6O0xah8s\nukgYRwoJRKq4nN2K6pog8n/nbUTKG7kHAgvQYVG7cECkkqv5+6qr7j/s5FMfi1OUeiwWrceK\ngUgFV0cLInVNELmhoZ+i5K/8k+x3fy/7U5zPrBeItCv6dScLInVNEL3HRj5Fyduu/H95BMxC\nWC8QqdjCxW5JpNYci1Mf6ilKvq1ic/UvsEogUs7Vji6S6dTH45IFiLQBIFK9kclWVJ9OJMOp\nj0+pGSJtAIhUb2TxHKlpgrTKeA1+NudIDzELYb1ApHoj1Kqd/tTH65IFVO02AESqN0IdR9Lj\nN40B40jrByLVG6HObNDjOR8ob+Qe/DZmIawZiFRvZKlqN0/8iXUGsP5JICBSvRE/kVK9ZAHz\nVkMBkeqN+ImU6CULiX6sLQKR6o34iZTof/qx/kkwIBIpxPKOTPKShWRP3bYHRCKFiH2YHIFI\nwYBIpBCxD5MjECkYEIkUIvZhciTVYuIGgUikEAx7OsaIDqp2wYBIpBD+OzpOWS/NYuIWgUik\nEN77OVbbkGQxcYtAJFII7/2MEZ2NA5FIIbz3s2P9DFPl1gJEIoXw3s9uIuEUZzVAJFII7/3s\nJBKKbusBIpFCeO/n6hzprcyqp4YTq/UAkUghvPdz2bj8N8ueGiYmrAeIRArhv6Or053/auqp\naasKEGk9QCRSCIY9/ehBlr31xNBT01cVINJ6gEikECz72uyFoaqAqXLrASKRQrDsa7NIhqoC\nqnbrASKRQrDsa7NIpr9gHGk1QCRSCJZ9be6pGRXDVLm1AJFIIVj2tbmnhqrC6oFIpBA8O9vY\nU0NVYfVAJFIIpr1t6qmhqrB6IBIphPRhQFVh7UAkUgjx44CqwsqBSKQQsQ8TE7i8SQyIRAoR\n+zDxgA6kHBCJFCL2YWIBJQ1BIBIpROzDxAIubxIEIpFCxD5MLGDYVxCIRAoR+zCxAJEEgUik\nELEPEwsQSRCIRAoR+zCxgIlIgkAkUojYh4kFVO0EgUikELEPEw8YR5IDIpFCxD5MTGAikhgQ\niRQi9mECqQORSCFiHyaQOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi\n9mECqQORSCFiHyaQOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mEC\nqQORSCFiHyaQOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQOR\nSCFiHyaQOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCFi\nHyaQOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCFiHyaQ\nOhCJFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCFiHyaQOhCJ\nFCL2YQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCFiHyaQOhCJFCL2\nYQKpA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCFiHyaQOhCJFCL2YQKp\nA5FIIWIfJpA6EIkUIvZhAqkDkUghYh8mkDoQiRQi9mECqQORSCGY9/qjb2XfesS8TRAViEQK\nwbvT38wK3uTdKIgKRCKFYN3nj7MH33vyvQfZY9atgqhAJFII1n3+MPtO/vM7aJK2BEQihQBg\nAemUEkns4UeSDxE6LjYdasvRsidBINLmN73KD706INLmN73KD706INLmN73KD706INLmN73K\nD706INLmN73KD706INLmN73KD706INLmN73KD706INLmN73KD706INLmN73KD706INLmN73K\nD706INLmN73KD706sCsAYAAiAcAARAKAAYgEAAMQCQAGIBIADEAkABiASAAwAJEAYAAiAcAA\nRAKAAYgEAAMQCQAGIBIADEAkABiASAAwEFSkiyN1dHHTC865wvnlcDvjWHybZvzUl8ejD8n2\nqcdb5vvQN+dKnV/1n+Hc1aslpEgn5bE8bh9fcYp0NdzOOBbfphk/9UW5oaMuB9k+9XjLjB/6\nqNxQzyTOXb1eAor0sjq62l0dqZebJ67UGdvG8+32qBcA/QAAA71JREFUv8okFt+m+T71lTq/\nKZq78+YJtk892TLfh74oNnrR2xznrl4xAUW6UPfzn/fU3eaJy+6fvlyqk0G2T2LxbZrvU59V\nm+22zvapJ1vm+9BH6mawZc5dvWYCinSmrneD/zZeqkuubauL3SDbJ7H4Ns34qesA7dYZP/Vo\ny9wfWh21/+T+0CsloEhq/J/JM3X/PD9N5dj21W6Y7ZNYfJtm/NQlN+qk+Sfjpx5tmflDX/S8\nZP7QayWuSCUnxne4bN4Qi2/TvJ+6aCzuj+Lw1THbLbN+6HtK9ZSESCUxRVLqXv6fzQumXkco\nkXg/9e76qOsT8X7q4ZYZP/Tl2VHvjAgilcQUqeKGqXIaSqQKrk99c9RrJVg/9WDL9VNsRerz\nTkmIVBLw6x8Z9rhEtptiMWx65jkHTvqpzfqpTzTSsGX7TVdt4N3VqyV41e56Ut6ROZHRx2LY\n9Mxz1lwfn1z3HjJ+6tGWa/iyfVxq5NrVqyWgSHfLc9/73YlqNSQhku2TWHybZvzU90dn/3yf\nerxlvg/dbKlt73h39WqJObPhotj5NxddbckLsZkNo03zferrcRWN7VNPtsz3ocuZDTdn3TkS\nZjaUhOzZHncl2DI1b6ppW0z/KWuyvfp9LFFZZ/7U56qbAcf7qSdbZtzVR6OjyLurV0tIkW7K\nacJVWNU8ccw15D4UqRdLYtMsn1pNROL61PotM+3qbksCu3q17HmtBQAeIBIADEAkABiASAAw\nAJEAYAAiAcAARAKAAYgEAAMQCQAGIBIADEAkABiASAAwAJEAYAAiAcAARAKAAYgEAAMQCQAG\nIBIADEAkABiASAAwAJEAYAAiAcAARAKAAYgEAAMQCQAGIBIADEAkABiASAAwAJEAYAAiAcAA\nRAKAAYgEAAMQCQAGIFJAzrubIZ8Vt2IF2wEiheS4uYfxZXdXcLAJIFJIrpS6Ln5fK3UV+7MA\nViBSUO5Wd/8+UXdjfxLAC0QKS6lQrRPYEBApLHmnbndTd/DAhoBIgblUdy+akgPYDhApNCcK\nHbsNApFCk3fsbmJ/BsAORAqOwj7fIDiowYFIWwQHNTgQaYvgoAYHIm0RHNTgQKQtgoMaHIi0\nRXBQAWAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkA\nBiASAAxAJAAYgEgAMACRAGAAIgHAAEQCgAGIBAADEAkABiASAAxAJAAYgEgAMACRAGAAIgHA\nAEQCgAGIBAADEAkABiASAAxAJAAY+P+vhgGFcM9kLgAAAABJRU5ErkJggg==",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model_cv, points_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred <- predict(svm_model_cv, points_data[, -3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction  -1   1\n",
       "        -1 122   8\n",
       "        1    3 117\n",
       "                                          \n",
       "               Accuracy : 0.956           \n",
       "                 95% CI : (0.9226, 0.9778)\n",
       "    No Information Rate : 0.5             \n",
       "    P-Value [Acc > NIR] : <2e-16          \n",
       "                                          \n",
       "                  Kappa : 0.912           \n",
       " Mcnemar's Test P-Value : 0.2278          \n",
       "                                          \n",
       "            Sensitivity : 0.9760          \n",
       "            Specificity : 0.9360          \n",
       "         Pos Pred Value : 0.9385          \n",
       "         Neg Pred Value : 0.9750          \n",
       "             Prevalence : 0.5000          \n",
       "         Detection Rate : 0.4880          \n",
       "   Detection Prevalence : 0.5200          \n",
       "      Balanced Accuracy : 0.9560          \n",
       "                                          \n",
       "       'Positive' Class : -1              \n",
       "                                          "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(pred, points_data$class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### But we should really be using the testing and training sets here. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = train_set, kernel = \"radial\", cost = 100, \n",
       "    gamma = 0.1, scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  radial \n",
       "       cost:  100 \n",
       "      gamma:  0.1 \n",
       "\n",
       "Number of Support Vectors:  28\n",
       "\n",
       " ( 15 13 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "\n",
    "svm_model_cv2 <- svm(class ~., \n",
    "                        data = train_set, \n",
    "                        kernel = \"radial\", \n",
    "                        cost = 100,\n",
    "                        gamma = 0.1,\n",
    "                        scale = FALSE)\n",
    "\n",
    "summary(svm_model_cv2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAABdFBMVEUAAAAOHBwPHx8QISES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\nABRAJAAFEAlAAUQCUACRABT4/5oroBtG4TsdAAAAAElFTkSuQmCC",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model_cv2, train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predTrain <- predict(svm_model_cv2, train_set[, -3])\n",
    "predTest <- predict(svm_model_cv2, test_set[, -3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 80  6\n",
       "        1   2 76\n",
       "                                          \n",
       "               Accuracy : 0.9512          \n",
       "                 95% CI : (0.9061, 0.9787)\n",
       "    No Information Rate : 0.5             \n",
       "    P-Value [Acc > NIR] : <2e-16          \n",
       "                                          \n",
       "                  Kappa : 0.9024          \n",
       " Mcnemar's Test P-Value : 0.2888          \n",
       "                                          \n",
       "            Sensitivity : 0.9756          \n",
       "            Specificity : 0.9268          \n",
       "         Pos Pred Value : 0.9302          \n",
       "         Neg Pred Value : 0.9744          \n",
       "             Prevalence : 0.5000          \n",
       "         Detection Rate : 0.4878          \n",
       "   Detection Prevalence : 0.5244          \n",
       "      Balanced Accuracy : 0.9512          \n",
       "                                          \n",
       "       'Positive' Class : -1              \n",
       "                                          "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTrain, train_set$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 42  2\n",
       "        1   1 41\n",
       "                                          \n",
       "               Accuracy : 0.9651          \n",
       "                 95% CI : (0.9014, 0.9927)\n",
       "    No Information Rate : 0.5             \n",
       "    P-Value [Acc > NIR] : <2e-16          \n",
       "                                          \n",
       "                  Kappa : 0.9302          \n",
       " Mcnemar's Test P-Value : 1               \n",
       "                                          \n",
       "            Sensitivity : 0.9767          \n",
       "            Specificity : 0.9535          \n",
       "         Pos Pred Value : 0.9545          \n",
       "         Neg Pred Value : 0.9762          \n",
       "             Prevalence : 0.5000          \n",
       "         Detection Rate : 0.4884          \n",
       "   Detection Prevalence : 0.5116          \n",
       "      Balanced Accuracy : 0.9651          \n",
       "                                          \n",
       "       'Positive' Class : -1              \n",
       "                                          "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTest, test_set$class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Last example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "points_data <- read.csv(\"datasets/data4.csv\", header = TRUE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th scope=col>X</th><th scope=col>Y</th><th scope=col>class</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><td>1.046319 </td><td>0.8279847</td><td>1        </td></tr>\n",
       "\t<tr><td>1.402901 </td><td>0.7013585</td><td>1        </td></tr>\n",
       "\t<tr><td>1.105754 </td><td>0.7297524</td><td>1        </td></tr>\n",
       "\t<tr><td>0.622754 </td><td>0.6807557</td><td>1        </td></tr>\n",
       "\t<tr><td>1.257972 </td><td>0.6321544</td><td>1        </td></tr>\n",
       "\t<tr><td>1.167278 </td><td>1.3824299</td><td>1        </td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|lll}\n",
       " X & Y & class\\\\\n",
       "\\hline\n",
       "\t 1.046319  & 0.8279847 & 1        \\\\\n",
       "\t 1.402901  & 0.7013585 & 1        \\\\\n",
       "\t 1.105754  & 0.7297524 & 1        \\\\\n",
       "\t 0.622754  & 0.6807557 & 1        \\\\\n",
       "\t 1.257972  & 0.6321544 & 1        \\\\\n",
       "\t 1.167278  & 1.3824299 & 1        \\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "X | Y | class | \n",
       "|---|---|---|---|---|---|\n",
       "| 1.046319  | 0.8279847 | 1         | \n",
       "| 1.402901  | 0.7013585 | 1         | \n",
       "| 1.105754  | 0.7297524 | 1         | \n",
       "| 0.622754  | 0.6807557 | 1         | \n",
       "| 1.257972  | 0.6321544 | 1         | \n",
       "| 1.167278  | 1.3824299 | 1         | \n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  X        Y         class\n",
       "1 1.046319 0.8279847 1    \n",
       "2 1.402901 0.7013585 1    \n",
       "3 1.105754 0.7297524 1    \n",
       "4 0.622754 0.6807557 1    \n",
       "5 1.257972 0.6321544 1    \n",
       "6 1.167278 1.3824299 1    "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "head(points_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "points_data$class <- factor(points_data$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {},
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAAP1BMVEUAAAAAv8QzMzNNTU1o\naGh8fHyMjIyampqnp6eysrK9vb3Hx8fQ0NDZ2dnh4eHp6enr6+vw8PDy8vL4dm3///92l2KZ\nAAAACXBIWXMAABJ0AAASdAHeZh94AAAgAElEQVR4nO2d62LbRpYGiWF8iSeX8Urv/6yrK0VK\n6CbQKADnoKt+ZD2xVKGbXy1FiKJPjyKymNPeN0DkCBiSCIAhiQAYkgiAIYkAGJIIgCGJABiS\nCEBbSP9ro/XzdOq8uIJiSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMY\nks5UTnj/GIakM5UT3j+GIelM5YT3j2FIOlM54f1jGJLOVE54/xiGpDOVE94/hiHpTOWE949h\nSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMYks5UTnj/GIakM5UT3j+G\nIelM5YT3j2FIOlM54f1jGJLOVE54/xiGpDOVE94/hiHpTOWE949hSDpTOeH9YxiSzlROeP8Y\nhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMYks5UTnj/GIak85NzWMEJuoJiSDpvnYMhtWBIOm+d\nA1+SIZUIcKA613EOhtSEIem8YVihJEMqEeBAda7iHAypDUPSec2wRkmGVCLAgers0wnvH8OQ\ndKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8fwxD\n0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEM\nSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8f4y2kETkBh+RdKZywvvHMCSd\nqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8fwxD0pnKCe8fw5B0\npnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jGyhvTw8IA7L+iM\n64T3j5E0pIeH95Ki3/E6DanM3gf68HApKfodr9OQyux9oIbUrRPeP4YhjaAzrhPeP0bOkHyO\n1K0T3j9G0pC8areHc1jBOdsVlKwh6dzeOQyNJRlSidZDAA9U59bOwZAqGJLOic6huSRDKtF6\nCOCB6tzYORhSDUPSOck5DO0lGVKJ1kMAD1Rnn054/xiGpDOVE94/hiHprDhbv3NUcy50BcWQ\nenFObOLG2XzBu+Jc6gqKIXXinJrE55CIkgypRIAD1Tk3pGlJXDvbL3iXnYtdQTGkPpyTk/gS\nElCSIZUIcKA65zinJ3HlXPKto5JzuSsohtSFsykkDEMqEeBAdc5wznhsif5nh/ePYUg6Uznh\n/WMYks5UTnj/GIakM5UT3j+GIelM5YT3j2FIOlM54f1jGJLOVE54/xiGpDOVE94/Rl8hXd7E\nC3RO5KDOzb/JC+8fo6uQPt5WknNO5ZjO7b/JC+8fo6eQrt7oGHNOJqRzpILZIU0oyZBKtB4C\neKANTkP6xFgEs5wTX3dkSCVaDwE80AanIX1irIL5Id0vyZBKtB4CeKAtTp8j3TAawRzn1NfC\nGlKJ1kMAD7TJ6VW7C0Ph4STa7fziCkpfIc1wTmxulpNmgXMwJBZDGndO/SpwjhNnUUiFr8uC\n3c6vrqAcMKTlBTxOvy4xw8nT7iw/s4l1O0dcQTleSEABhw+pco0g1O0ccwXlcCERCRgSiyGV\nCHCgJScT0rGfI9UuW0e6naOuoBhSwXnwq3ZpnfD+MQ4XEvMciUcn5ArKHiGh/89+nat2PDoh\nV1B2CIl9+hH9jtdpSGWWHCh8QSz6Ha/TkMosOVBD0rnEFRRD0pnKCe8fw+dIOlM54f1jHPCq\nnc4jO+H9Y6wa0udiot9JOuM74f1jrBnSl6/hot9JOuM74f1jrBjS16sK0e8knfGd8P4xDEln\nKie8fwxD0pnKCe8fI81zpNKlvuh3vE5DKjPxDw1etSt+8yn6Ha/TkMpsfqDll0NEv+N1GlKZ\nzQ/UkHS+uYJiSDpTOeH9YyQJyedIOt9cQckSklftdL66gpImJJ06X1xBMSSdmLP011IYUonW\nQwAPVGc45yZvgwzvH8OQdFLOTd69Fd4/hiHphJzbvA0yvH8MQ9IJObd5G2R4/xiGpJNxVt5Q\n3JBKtB4CeKA6+3TC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEM\nSWcqJ7x/DEPq3Pn5+z5Rb+fFFRRD6tv55TuoQW/nhysohtS305AgDKlr59cX9cS8nVeuoBhS\n105DojCknp0jrzMNeTuvXUExJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQunGW3uNniXMahlQi\nwIHqnOksvsfPAudEDOma89WvVzjQ1r/qPPodH8ZpSKsyOaTzuiEV39t7gbOZi7O17poTZI6z\n/B4/7c6pGNIH53Ufkcp/20S7s513Z3PdFSfJ7JAmlLT37bzrCkqQL+1ChtR+o8pOlBnO8lv8\ntDsnY0jXvIX0n2fwm/G2Wdy7iJA3qpH3N8va+3YcliCPSCGfIx3pEWmBc9pl83nOBa6gRAkp\n5FW7Iz1HanMOk6/2TXcudAUlTEiteNVuRefkZ1YznEtdQTEknWVn5W2Im51LXUExJJ1F52BI\nkzEknUXnAJVkSCUCHKjO1Z3QM6QnJ+C4uIJiSDpLTugLu2fRYsUFeP8YhqRzdSfysPYGvH+M\nkCHNueo8xTn3KnaWgSZxMl8fvgHvHyNiSLO+DzrBOfv7qkkGupVzaQXQM61X4P1jBAxp3itz\n7jvnv9In8ehXcC6tgLpk8Qq8fwxDanLOJ6+TCYkqCd4/hiE1OeeT1un3kSYRMCSfI4VyAo8n\nhlRi5QP1ql0cJ/GFmSGVCHCgKzgvvQW/nZs6ie/KGlKJAAfKOz++Aox9O/t2wvvHMKR3rq5J\nhL6dnTvh/WMY0juGlMIJ7x/DkN4xpBROeP8YEUJa9APdPkfqywnvHyNASMveYsSrdn054f1j\n7B/Swje9in7H6zSkMuSB1kKaEFj0O16nIZUhD7QS0utv1GuKfsfrNKQy6IHWO7r3hV/0O16n\nIZVhD7Tyhd39kqLf8ToNqcw2B2pIqzrHXz4X73Z+cgUlckiTvraLfseHdRZehxrudn52BSV0\nSP97u9awynOkijX6mBBn4RXd4W7nZ1dQYof0QnHxL7/ReltqfUYfE+Es/WjEmxN8JzpDKhPg\nQN+/8Gt0Vr9ijDb6FZzFHzJ6fP/tdnfBCbmCkjek+1fGJ3z2+G8GG/0aTkOCyR9SW0mdh1T+\nudfHy283y8edlCsovYbU/XOkqpN9KzpDKhLgQJd+bdf9VbuKk30nOkMqE+BA//deEut8IdPo\n13Bifw3FlRNzBSVzSC8pZRqoTsAVlNwh6ezOCe8f46ghXX/NN/vrv+hj6tkJ7x/joCFdP3ua\n/0wq+ph6dsL7xzhmSNfX8xqu7UUfU89OeP8YhjTVuRCdkCsonYU0rajoY+rZCe8f45ghlZ4j\nTXxsij6mnp3w/jEOGtL4VbupX+VFH1PPTnj/GEFDmv6kZs5tMaT8Tnj/GDFDmnF5wJD6csL7\nxwgZ0pwLbbNui8+R0jvh/WP0FZJX7dI74f1j9BGS30c6jBPeP0bIkOjnSHO/Ixt9TJs6gR+m\nMKQSqx8oetVu9msb0o5+BSfxY0mGVCLAgc5wGtICpyFNw5DanLPJ5bzEg/ykrCGVCHCgc5w+\nR5rp/IgH+ZlzQyoR4EBnOb1qN895aYd58wZDKhHgQHWu5/yIh3kbFEMqEeBAda4cUns8Xz7T\nkEq0HgJ4oDpXcy58FPr6qYZUovUQwAPVuXZIrSV9/VRDKtF6COCB6gzqHGnQkEq0HgJ4oDrH\nnbtfYRt5NDOkEq2HAB6ozlHn7lfYxr4uNKQSrYcAHqjOdUJC/6q+NwypRIAD1TnmXPxNH/Zv\nGHvDkEoEOFCdxZCWpGBIjRjSkZyLX4cA/1V9bxhSiQAHqrMc0oIXJaxSkiGVCHCgOldwzuxw\n6gcaUonWQwAPVOd6IU0MZPIHGlKJ1kMAD1Tn7k5DusKQdDY6pz92GVKJ1kMAD1Tn3s7pXwUa\nUonWQ5j9Gfd/tPV4A03inPF0ypBKtB7C3E+Y8GYL6Jje/muHG/0azhkXJgypROshzPz4KW//\nQ95J7/+1w41+DeeMy3uGVKL1EGZ+/MYhXf5zhxv9Cs4518kNqUTrIcz8eEMinewVtjnfcTKk\nEq2HMPcTtn2OdOyQdrgwsIYT3j9G7JA2vmp36OdIhrQqwUPa2Hngq3Z7XGFbwwnvH8OQOnFO\nfUqz9+286wqKIfXhnHxtIPqfHd4/hiH14Zz87dPof3Z4/xiGpDOVE94/hiHpTOWE949hSDpT\nOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMYks5UTnj/GIakM5UT3j+GIelM\n5YT3j2FIOlM54f1jGJLOBU5/QvYdQ9LZ7vTtuC4Yks52p2/HdcGQdDY7fV+7DwxJZ7Nzjx8W\nhPePYUg6W527/LAgvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jHaQhKRG3xE0pnKCe8fw5B0\npnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPS\nmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8fwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJ\nZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAk\nnamc8P4xDElnKie8fwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQ\ndKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8fwxD\n0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEM\nSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8fwxD0pnKCe8fw5B0pnLC+8cw\nJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/D\nkHSmcsL7xzAknamc8P4xDElnKie8fwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8M\nQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4x\nDElnKie8fwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvH\nMCSdqZzVXf777XT6Y+Tf/zqPf/xfP65KqKfw4696EoakM5Wzusvz6TTaQyGSv8/3P+bD/Xf1\ntw1JZypnfc2FORf+/fnXhM99p/So9v7p9c8uEOBAdfbprI759QHp7++n0/nn87/49/vLL94e\np/79cTr9+Pfl4/45f3t8/Hm++pjXkD4+9c/z6Y9fN794/feGpPMQzvsh/fXyf05Ps//98qXe\n97eQXv/X+ffzx307/Xj8/fwhl495CenjU3++/OLX1S+efvXbkHQexVnd5cujyh+n/z4+/vP8\ny59Pufz9/IuXf//z9PQo9O308gj1nNCfp78eP33Mx6eeTv8+/evz1S+eKvvTkHQexXk/pKev\n1v7689trF79vAnv6su7f56t6z3E8Pn5/+eftx1w+9Xx6u0x3+cXTp343JJ1HcU4I6dvp7cnS\n5frByy9e/9fHv7/+5/svLp/619NXfH88h3b5xZ3LEYakM5Xzfkg/Tn/8+uvfppA+PvXpK7w/\nTq9XvC+/MCSdx3HeD+nlH7/rX9p9dHHzMR+f+syvT78wJJ3HcVZ3+RbS34+/v71ebPj5fung\n8fZiw/MHvz5HuvmYj089P/3qn+drDJdf+BxJ55Gc90N6vWD9/Mt/Xy5tvzwEnW8vfz9/8OtV\nu6uPuf7U11/9efULr9rpPJLzfkhPz3RO316uaD/+8+31W7C/Xh9SPr4h+/xhr99HunzM2/Or\n90/9eT6dX8L5+IXfR9J5HGfTXsf5OW/8J1/ZoPM4zqa9Frh5rd09fK2dziM5m/Za4O96G7f4\n6m+dR3I27bXE9c8j3cGfR9J5KGfTXjfAkHSmcsL7xzAknamc8P4xDElnKie8fwxD0pnKWd3l\n/1VoGvp0DElnKmd1l4bUjs6+nNVdJgjp/MTH/wpwoDr7dFZXGj+k8+UfLwQ4UJ19OqszNaR2\ndPblrM7UkNrR2ZezOtNMIf3nmfVuj0g7mUJ6IcD/Z9LZp7M6U0NqR2dfzupMDakdnX05qzOd\nGlLbd0+rXCsrP5xhSDpjOKtrvg7n4eGhFNK9v3iihWvlqfxDgIakM4azuubbjm5Lut78yiF9\nP52KD0q+skFnCGd1zZ86uimptHqIG+V/z5UHpWsCHKjOPp3VXUYJ6fF37UHpigAHqrNPZ3WX\nYUJ6fPzn++XNJisEOFCdfTqru5z2HMmQdOqs7vLOVbvLrv3STmfvzuou/69CefUIXmzQmcpZ\n3WWQkCY+HD0aks7dnNVdBglp4sPRoyHp3M1Z3eXUkFZg4kuEPhHgQHX26azuMkhI0wlwoDr7\ndFZ3aUjt6OzLWd2lIbWjsy9ndZeG1I7OvpzVXRpSOzr7clZ3aUjt6OzLCe8fw5B0pnJWd+kj\nUjs6+3JWd2lI7ejsy1ndpSG1o7MvZ3WXhtSOzr6c1V0aUjs6+3JWd2lI7ejsy1ndpSG1o7Mv\nZ3WXhtSOzr6c1V0aUjs6+3JWd3lTzjAa0go/HLvAG+BAdfbprO7ypqNhGAlpjbf9fhU3fVaA\nA9XZp7O6y7shrfK232/mFgIcqM4+ndVd3nZ0W9Kiwd/HkHSmclZ3+TmkwZB06hx31fjckSHp\n1Flw1fgS0mBIOnWOumr8X4VFg7+PIelM5azu0pDa0dmXs7pLQ2pHZ1/O6i4NqR2dfTmru5wQ\n0loYks5UzuouDakdnX05q7s0pHZ09uWs7tKQ2tHZl7O6S0NqR2dfzuouDakdnX05q7s0pHZ0\n9uWs7tKQ2tHZlxPeP4Yh6UzlhPePYUg6Uznh/WMYks5UTnj/GIakM5UT3j+GIelM5YT3j2FI\nOlM54f1jGJLOVE54/xiGpDOVE94/hiHpTOWE949hSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh\n6UzlhPePYUg6Uznh/WMYks5UTnj/GIakM5UT3j+GIelM5YT3j2FIOlM54f1jGJLOVE54/xiG\npDOVE94/hiHpTOWE949hSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMY\nks5UTnj/GIakM5UT3j+GIelM5YT3j2FIOlM54f1jGJLOVE54/xiGpDOVE94/hiHpTOWE949h\nSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMkCmlYwTmOzrhOeP8YeUIa\nhtGSot/xOg2pzB4Hakg6/2dIiw90GMZLin7H6zSkMmsc6PhToI/fLpQU/Y7XaUhlVjjQwldu\nV79tSDoN6d6BFr5wu/n90Y+JfsfrNKQy/IGWngItcTajM64T3j9GkJBKX7gtcbajM64T3j9G\njJCKX7gtcC5AZ1wnvH+MGCEtQGdfTnj/GIakM5UT3j+GIelM5YT3j3HwkJquX4QfU89OeP8Y\nxw7JK4GHc8L7xzh8SF4JPJYT3j/GoUPykvp2zoeHB9w56grK8UNqKCnSQJM4Hx4qJRlSbt6/\nzbv37eiAh1f2vhk7cuhHJJ1LnbUv2D59XOUhyUekEq2HwJ3n5Qu2nANN4qx+wfblAw1pNq2H\ngB3nx1OflANN4qzncYPPkZo+q/UQsOM0pC2cM0Lyql3TZ7UeAnWaV1fjMg40i3NOSFOdi11B\nMaQRdL7BdGRIRXY+0OufXko50DROpCNDKhLgQHX26YT3j2FIOlM54f1jGJLOVE54/xiGpDOV\nE94/hiHpTOWE949hSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg613FevSjCkEq0\nHgJ4oDpjO69fpmdIJVoPATxQnaGdNy8cN6QSrYcAHqjO0E5DmkTrIYAHqjO005Am0XoI4IHq\njO30OdIUWg8BPFCdwZ1etZtA6yGAB9qr8/pH7SLfzrWc8P4xDCmX8+aHvwPfztWc8P4xDCmV\n8/btSOLezvWc8P4xDCmV05Dg/WMYUiqnIcH7xzCkXE6fIwXFkJI5vWoXE0PSmcoJ7x/DkHSm\ncsL7xzAknamc8P4xDKk7Z/1tiOPczoIrKIbUm/POG+OHuZ0lV1AMqTPnvb+qJcrtLLqCYkid\nOQ1pHQypM6chrYMh9eb0OdIqGFJ3Tq/arYEh6UzlhPePYUg6Uznh/WMYks5UTnj/GIakc46z\n6W9nNqQSrYcAHqjOPZx3Lvk1OWe6gmJIOqc7730TqsU51xUUQ9I53WlIRQxJ53SnIRUxJJ0z\nnD5HKnHAkIYVnMs5iHOko/tpGVKJ1kMAD7ToHIbFJR1k9Ns4JzxIGVKJ1kMAD7ToNKRNnVOe\nNhlSidZDAA+05ByG5SVFGGgWpyG9YkgTnARHdRrSKzFDmlHCJ+cwACVFGOgqzgmX3HyO1EbA\nkIZZT3NCDDSJc8rF6/m306t2z8QLaeZDSoSBJnFO+nZqgNtZdwUlakiTS4p+xwdyGtKKhAtp\nMKS1nIa0ImFDmlpS9Ds+knOd50j3MaQS6x3o3Aek8Hd8KOcaV+0mYEglAhyozj6d8P4xDEln\nKie8f4xUIY19wRf9jtdpSGWqf9Ty85uFBzr61Cn6Ha/TkMrU/qSVKwXHCmne65CiDzSLE94/\nBh5S7ZLb3AO9FY1fzdvtjp/5gr7oA83ihPePsUpIpYnNPNBPovHr4juGNKuk6APN4oT3j0GH\nVP0u0PyQhtv/FSikua8xjz7QLE54/xhwSPXvp8470GE0pM/qne742T+tEX2gWZzw/jECX/6e\ntlVD6ssJ7x8jbkgTXyy0zx0/+5VM4QeaxQnvHyNuSDp1jrmCYkg6Uznh/WMYks5UTnj/GIak\nM5UT3j/G8UJa/o7F4cfUsxPeP8bhQlr+Pqvxx9SzE94/xgFD8k30R5j8l0hE/7PD+8c4WkjA\n+6zGH9N85/S/jiX6nx3eP8bBQiLeZzX+mGY7Z/wFYdH/7PD+MQzprpPBkCBXUI4V0vxX7tx3\nQhgS5ApKrJAaCoh+x8dw+hxpbUKF1PJYEv2OD+L0qt3KGJLOVE54/xiRQmp6ehP9jtdpSGXW\nOdCmCwXR73idhlRmlQNtu+IW/Y7XaUhlVjnQtmvX0e/4fZyTry3McLZjSCUCHKjOinP61e7p\nzgUYUokAB6qz7Jzx/dfJziUYUokAB6rTkCJhSAd0GtL2GNIRnT5H2hxDOqTTq3ZbY0g6Uznh\n/WMYks53Xh7Got9OeP8YXYR07FdLtH8Zd8vrE6vof3Z4/xg9hHTsV0ssuLBww9ulvuh/dnj/\nGJ2EdNy/XW/Jpe4bDGkRHYQ04QV8n347+piunYYUg+khna9+HeBApzvvvxT2829HH9O1EwvJ\n50iLmBzSOWtI919U/uV3o4/pxkl15FW7RUwN6Zz3EekuXzoLejsLTqqjayeLIV1z2JC+PmDF\nvJ06X1xBmR3Sf55Z6cbswvuXfnvfDkmNj0g6Uzn5BBgMSWcqJ58Aw92Qzue363WGpDOAc60Q\nlpLzEenq0kD0O16nId0SKKTri2zR73idhnSLIekM4OQTYMj4Wrub7/tEv+N1GlKZfQ/05juo\n0e94nYZUZtcDHQypYye8f4y8IQ2k8xadcZ3w/jEShqRze+eyF8YaUokAB6pzQ+fCH9UwpBIB\nDlTnds6lPzxoSCUCHKjOGc4ZFXxxPlyYqrjvXAC8fwxD6sA5J4PPzgdDmoQhHd85q4NPTqIj\nQyoS4EB1bhzSvJtWcy4C3j+GIR3fCYQ073bdcS5zBcWQOnAufo404zZNcS5yBcWQenAuvGo3\n9XMnO5e4gmJIOlM54f1jGJLOVE54/xiGpDOVE94/hiH16xx99hPwdt66gmJI3TrHr8fFu52f\nXEExpF6dhe8Qhbudn11BMaRenYaEYki9Og0JxZC6dfocicSQ+nV61Q7EkHTecfpauykYks66\n01d/T8KQdFadxM9RGFKJAAeq86gh1f9r8P4xDEln1bl5SHf+c/D+MQxJZ9258XOke+HC+8cw\nJJ13nNtetTMk+EB19uk0JPhAKwzX/2PvO14n7PQ5EnygRQZDOrTTq3bwgY7x3NBwU9L+d7zO\nLZ3w/jFyhfSc0GBIPTvh/WNkC2m4/Rtk49/xOg2pzE4HOgy3f1sf4RxDZ1wnvH+MTCENw0hJ\n0e94nYZUZp8D/fxgRDhH0RnXCe8fI1FIX7+sW+4cR2dcJ7x/jEQh6dRpSPiB6uzTCe8fw5B0\npnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPS\nmcoJ7x8je0hD9DtepyGVCXCgr3x9EStB9DH17IT3j2FII0QfU89OeP8YuUMa+7kKgOhj6tkJ\n7x/jACHdlIRkFX1MPTvh/WOkDunrDygxj0/Rx9SzE94/xhFCGm7+DeC9dztb/hvRB5rFCe8f\nI1lIXyd844SeMt25nU3/kegDzeKE94+RK6SRBX8JCSjJkOI64f1jZAvpy4Qfb397g5Da/iPR\nB5rFCe8fI1VIYwv+GtLykgwprhPeP0amkGK8HVdjrdEHmsUJ7x/DkEbQGdcJ7x8jUUhT39du\n7S/tdF6Y+Jf5GVKJ1kMAD7ToXP05ks53pv71soZUovUQwAMtOoGrDXsPNJizVMvkv/DckEq0\nHgJ4oCUncdku8ehXcBZrMaQrjhYScgE87+jvcX/3X5zlXAzpCkO670QI4Zww/Bkh+RzpioOF\nNH5lb5mTIYJzykPInOrx1EoAAAVuSURBVJC8avfBwULSWaMppMmPO3OcC1xBMaSOnG0hTX3c\nmeVsdwXFkHpytjxHAjCkEgEOVOdGV+0ADKlEgAPV2acT3j+GIelM5YT3j2FIOlM54f1jGJLO\nVE54/xiGpDOVE94/hiHpTOWE94+RL6S7P9i3HJ1xnfD+MdKFdP8nZJejM64T3j9GwpBuS4p+\nx+s0pDL7HeiXl3ZHv+N1GlKZ3Q706w9JRL/jdRpSmd0O1JB6d8L7x8gV0sjP7UW/43UaUpkA\nB6ozkPPOi8oNqUTrIYAHqjOO896PORlSidZDAA9UZxjn3R+8NaQSrYcAHqjOME5DejQkncud\nhvRoSDoBp8+RDEkn4fSqnSHpzOWE949hSDpTOeH9YxjS8ZyL3tAx+p8d3j+GIR3OuewthqP/\n2eH9YxjS0ZyT/7KVGc6FGJLk4y2kvW9Gb/iIdDSnj0i7YEiHc/ocaQ8M6XhOr9rtgCHpTOWE\n949hSDpTOeH9YxiSzlROeP8YhqQzlRPeP4Yh6UzlhPePYUg6Uznh/WMYks5UTnj/GIakM5UT\n3j+GIelM5YT3j2FIOlM54f1jGJLOVE54/xiGpLPZOfVFfYZUovUQwAPVubtz8svMDalE6yGA\nB6pzb+f0H3wypBKthwAeqM69nYZ0jSHpbHQa0jWGpLPV6XOkKwxJZ7PTq3YfGJLOVE54/xiG\npDOVE94/hiHpTOWE949hSDpTOeH9YxiSzpnO+e/2ZUglWg8BPFCdOzkb3n/SkEq0HgJ4oDr3\ncba8I7IhlWg9BPBAdRpSJAxJ5yynIY1jSDrnOX2ONIoh6Zzp9KrdGIakM5UT3j+GIelM5YT3\nj2FIOlM54f1jGJLOVE54/xiGpDOVE94/hiHpTOWE949xzJCGFZwL0Qm5gnLIkIZhWUnRx9Sz\nE94/xlFDWlRS9DH17IT3j3HEkAZDOq4T3j/GYUNaUlL0MfXshPePccCQhmFpSdHH1LMT3j/G\nAUPSeWQnvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7\nxzAknamc8P4xDElnKie8fwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnv\nH8OQdKZywvvHMCSdqZzw/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8\nfwxD0pnKCe8fw5B0pnLC+8cwJJ2pnPD+MQxJZyonvH8MQ9KZygnvH8OQdKZywvvHMCSdqZzw\n/jEMSWcqJ7x/DEPSmcoJ7x/DkHSmcsL7xzAknamc8P4xDElnKie8f4y2kALxn71vwES8ncfG\nkDbC23lsDGkjvJ3HxpA2wtt5bAxpI7ydxyZ9SCIRMCQRAEMSATAkEQBDEgE4QEjnvW/ABM5P\n7H0bppHkZoYjf0gZFnq+/CM6GU4zJOlDOmcYaJqQUpxmSNKHlGKgaULKcivjYUhbYEiHx5C2\nwJAOjyFtgSEdnrwhXa4oJ7jrDenw5A3pQoK73pAOjyFtgSEdHkPaBF/ZcHQOEJLI/hiSCIAh\niQAYkgiAIYkAGJIIgCGJABiSCIAhiQAYkgiAIW3Jj9P3p39+O/3Y+4YIjSFtye/z6Z/HX6fz\n771viNAY0qb8On1/iunX3jdDcAxpW56+rDv9sfeNEB5D2pZ/Tqenr+7kcBjSxvzwSsMhMaSN\nOZ/80bkjYkjb8vP0/fRz7xshPIa0KU9Pkf7ySdIRMaRN+ePpGdKP07e9b4bgGNKWvHwv1m8k\nHRFD2pCnhP779H/+60sbjochbcj7F3W+2O54GJIIgCGJABiSCIAhiQAYkgiAIYkAGJIIgCGJ\nABiSCMD/Az2BStGCwyx6AAAAAElFTkSuQmCC",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot1 <- ggplot(points_data, aes(X, Y)) + \n",
    "            geom_point(aes(color = factor(class), shape = factor(class)))\n",
    "\n",
    "plot1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tctrl <- tune.control(sampling = \"cross\", \n",
    "                         cross = 10, \n",
    "                         best.model = TRUE,\n",
    "                         performances = TRUE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Parameter tuning of 'svm':\n",
       "\n",
       "- sampling method: 10-fold cross validation \n",
       "\n",
       "- best parameters:\n",
       " gamma cost\n",
       "   0.1    1\n",
       "\n",
       "- best performance: 0 \n",
       "\n",
       "- Detailed performance results:\n",
       "   gamma cost error dispersion\n",
       "1  1e-06    1  0.61 0.05676462\n",
       "2  1e-05    1  0.61 0.05676462\n",
       "3  1e-04    1  0.61 0.05676462\n",
       "4  1e-03    1  0.61 0.05676462\n",
       "5  1e-02    1  0.53 0.17669811\n",
       "6  1e-01    1  0.00 0.00000000\n",
       "7  1e-06   10  0.61 0.05676462\n",
       "8  1e-05   10  0.61 0.05676462\n",
       "9  1e-04   10  0.61 0.05676462\n",
       "10 1e-03   10  0.61 0.05676462\n",
       "11 1e-02   10  0.40 0.14907120\n",
       "12 1e-01   10  0.00 0.00000000\n",
       "13 1e-06  100  0.61 0.05676462\n",
       "14 1e-05  100  0.61 0.05676462\n",
       "15 1e-04  100  0.61 0.05676462\n",
       "16 1e-03  100  0.60 0.08164966\n",
       "17 1e-02  100  0.00 0.00000000\n",
       "18 1e-01  100  0.00 0.00000000\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tuned_params_cv <- tune(svm, class ~. ,\n",
    "                           data = points_data,\n",
    "                           kernel = \"radial\",\n",
    "                           ranges = list(gamma = 10^(-6:-1), cost=10^(0:2)), \n",
    "                           tunecontrol = tctrl)\n",
    "\n",
    "summary(tuned_params_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "split <- 0.65\n",
    "\n",
    "trainIndex <- createDataPartition(points_data$class, p = split, list = FALSE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>66</li>\n",
       "\t<li>3</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 66\n",
       "\\item 3\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 66\n",
       "2. 3\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] 66  3"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_set <- points_data[trainIndex, ]\n",
    "dim(train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>34</li>\n",
       "\t<li>3</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 34\n",
       "\\item 3\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 34\n",
       "2. 3\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] 34  3"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_set <- points_data[-trainIndex, ]\n",
    "dim(test_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\n",
       "Call:\n",
       "svm(formula = class ~ ., data = train_set, kernel = \"radial\", gamme = 0.1, \n",
       "    cost = 100, scale = FALSE)\n",
       "\n",
       "\n",
       "Parameters:\n",
       "   SVM-Type:  C-classification \n",
       " SVM-Kernel:  radial \n",
       "       cost:  100 \n",
       "      gamma:  0.5 \n",
       "\n",
       "Number of Support Vectors:  7\n",
       "\n",
       " ( 4 3 )\n",
       "\n",
       "\n",
       "Number of Classes:  2 \n",
       "\n",
       "Levels: \n",
       " -1 1\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_model <- svm(class ~.,\n",
    "                    data = train_set,\n",
    "                    kernel = \"radial\",\n",
    "                    gamme = 0.1,\n",
    "                    cost = 100, \n",
    "                    scale = FALSE)\n",
    "\n",
    "summary(svm_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0gAAANICAMAAADKOT/pAAAAsVBMVEUAAAAfPj4fPz8mTU0w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\nRSJEgf8PDla0ex3itr8AAAAASUVORK5CYII=",
      "text/plain": [
       "Plot with title \"SVM classification plot\""
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot(svm_model, train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predTrain <- predict(svm_model, train_set[, -3])\n",
    "predTest <- predict(svm_model, test_set[, -3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 33  0\n",
       "        1   0 33\n",
       "                                     \n",
       "               Accuracy : 1          \n",
       "                 95% CI : (0.9456, 1)\n",
       "    No Information Rate : 0.5        \n",
       "    P-Value [Acc > NIR] : < 2.2e-16  \n",
       "                                     \n",
       "                  Kappa : 1          \n",
       " Mcnemar's Test P-Value : NA         \n",
       "                                     \n",
       "            Sensitivity : 1.0        \n",
       "            Specificity : 1.0        \n",
       "         Pos Pred Value : 1.0        \n",
       "         Neg Pred Value : 1.0        \n",
       "             Prevalence : 0.5        \n",
       "         Detection Rate : 0.5        \n",
       "   Detection Prevalence : 0.5        \n",
       "      Balanced Accuracy : 1.0        \n",
       "                                     \n",
       "       'Positive' Class : -1         \n",
       "                                     "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTrain, train_set$class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Confusion Matrix and Statistics\n",
       "\n",
       "          Reference\n",
       "Prediction -1  1\n",
       "        -1 17  0\n",
       "        1   0 17\n",
       "                                     \n",
       "               Accuracy : 1          \n",
       "                 95% CI : (0.8972, 1)\n",
       "    No Information Rate : 0.5        \n",
       "    P-Value [Acc > NIR] : 5.821e-11  \n",
       "                                     \n",
       "                  Kappa : 1          \n",
       " Mcnemar's Test P-Value : NA         \n",
       "                                     \n",
       "            Sensitivity : 1.0        \n",
       "            Specificity : 1.0        \n",
       "         Pos Pred Value : 1.0        \n",
       "         Neg Pred Value : 1.0        \n",
       "             Prevalence : 0.5        \n",
       "         Detection Rate : 0.5        \n",
       "   Detection Prevalence : 0.5        \n",
       "      Balanced Accuracy : 1.0        \n",
       "                                     \n",
       "       'Positive' Class : -1         \n",
       "                                     "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusionMatrix(predTest, test_set$class)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "R",
   "language": "R",
   "name": "ir"
  },
  "language_info": {
   "codemirror_mode": "r",
   "file_extension": ".r",
   "mimetype": "text/x-r-source",
   "name": "R",
   "pygments_lexer": "r",
   "version": "3.4.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}