Assignment2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pylab as plt\n",
    "import pandas as pd\n",
    "import sys\n",
    "\n",
    "def sigmoid(x):\n",
    "    return 1/(1+np.exp(-x))\n",
    "\n",
    "\n",
    "def w_calculator(Features,y,eta):\n",
    "    w = np.mat('[0;0;0;0]')\n",
    "    for i in range(10000):\n",
    "        pr = sigmoid(Features * w)\n",
    "        w = w + eta * Features.T * (y - pr)    \n",
    "    return w\n",
    "\n",
    "def test(w_n):\n",
    "    print(\"-------results for values of 1--------\")\n",
    "    for i in range (50):\n",
    "        a = sl[i]\n",
    "        b = sw[i]\n",
    "        c = pl[i]\n",
    "        d = pw[i]\n",
    "        Feature = np.hstack([a,b,c,d])\n",
    "        print (sigmoid(Feature*w_n))\n",
    "    print(\"-------results for values of 2--------\")\n",
    "    for i in range (51,100):\n",
    "        a = sl[i]\n",
    "        b = sw[i]\n",
    "        c = pl[i]\n",
    "        d = pw[i]\n",
    "        Feature = np.hstack([a,b,c,d])\n",
    "        print (sigmoid(Feature*w_n))\n",
    "    print (\"-------results for values of 3--------\")\n",
    "    for i in range (101,150):\n",
    "        a = sl[i]\n",
    "        b = sw[i]\n",
    "        c = pl[i]\n",
    "        d = pw[i]\n",
    "        Feature = np.hstack([a,b,c,d])\n",
    "        print (sigmoid(Feature*w_n))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df_iris = pd.read_csv(u'data/iris.txt',sep=' ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sl = np.matrix(df_iris.sl[:]).T\n",
    "sw = np.matrix(df_iris.sw[:]).T\n",
    "pl = np.matrix(df_iris.pl[:]).T\n",
    "pw = np.matrix(df_iris.pw[:]).T\n",
    "FeatureVector = np.hstack((sl,sw,pl,pw))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classify 1 versus 2,3\n",
    "### In this case 2 and 3 will be interpreted as 0 as logistic regression is binary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.93933138]\n",
      " [ 3.02491508]\n",
      " [-4.64943993]\n",
      " [-2.17020562]]\n"
     ]
    }
   ],
   "source": [
    "y = np.matrix(df_iris.c[:]).T\n",
    "y[y > 1] = 0\n",
    "w_1 = w_calculator(FeatureVector,y,0.001)\n",
    "print(w_1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## Test w_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------results for values of 1--------\n",
      "[[ 0.99978281]]\n",
      "[[ 0.9988119]]\n",
      "[[ 0.99950787]]\n",
      "[[ 0.9981486]]\n",
      "[[ 0.99982369]]\n",
      "[[ 0.99969577]]\n",
      "[[ 0.99941619]]\n",
      "[[ 0.9994862]]\n",
      "[[ 0.997432]]\n",
      "[[ 0.99887493]]\n",
      "[[ 0.99985755]]\n",
      "[[ 0.9990135]]\n",
      "[[ 0.99894935]]\n",
      "[[ 0.99958319]]\n",
      "[[ 0.99999021]]\n",
      "[[ 0.99998004]]\n",
      "[[ 0.99995262]]\n",
      "[[ 0.99973019]]\n",
      "[[ 0.99974999]]\n",
      "[[ 0.99982665]]\n",
      "[[ 0.99910609]]\n",
      "[[ 0.9997086]]\n",
      "[[ 0.99996002]]\n",
      "[[ 0.99693201]]\n",
      "[[ 0.99603203]]\n",
      "[[ 0.99726329]]\n",
      "[[ 0.99873849]]\n",
      "[[ 0.99968528]]\n",
      "[[ 0.99973246]]\n",
      "[[ 0.99801756]]\n",
      "[[ 0.99755895]]\n",
      "[[ 0.99945537]]\n",
      "[[ 0.99995874]]\n",
      "[[ 0.99998205]]\n",
      "[[ 0.99887493]]\n",
      "[[ 0.99976671]]\n",
      "[[ 0.99990629]]\n",
      "[[ 0.99887493]]\n",
      "[[ 0.99880628]]\n",
      "[[ 0.99953225]]\n",
      "[[ 0.9998138]]\n",
      "[[ 0.98889247]]\n",
      "[[ 0.99934778]]\n",
      "[[ 0.99856138]]\n",
      "[[ 0.99861856]]\n",
      "[[ 0.99837927]]\n",
      "[[ 0.99977789]]\n",
      "[[ 0.99913972]]\n",
      "[[ 0.99984352]]\n",
      "[[ 0.99956321]]\n",
      "-------results for values of 2--------\n",
      "[[ 0.00020622]]\n",
      "[[  3.79599535e-05]]\n",
      "[[  9.18373373e-05]]\n",
      "[[  4.24425828e-05]]\n",
      "[[  4.91887746e-05]]\n",
      "[[  8.07001648e-05]]\n",
      "[[ 0.00350226]]\n",
      "[[  9.73743671e-05]]\n",
      "[[ 0.00029765]]\n",
      "[[ 0.00045409]]\n",
      "[[ 0.00028403]]\n",
      "[[ 0.00020812]]\n",
      "[[  3.07841318e-05]]\n",
      "[[ 0.00396316]]\n",
      "[[ 0.00039943]]\n",
      "[[  5.31262021e-05]]\n",
      "[[ 0.00049152]]\n",
      "[[  8.30104862e-06]]\n",
      "[[ 0.00045376]]\n",
      "[[  1.66697463e-05]]\n",
      "[[ 0.00073175]]\n",
      "[[  3.51844214e-06]]\n",
      "[[  3.51120594e-05]]\n",
      "[[ 0.00032548]]\n",
      "[[ 0.00026874]]\n",
      "[[  2.75802248e-05]]\n",
      "[[  9.46137145e-06]]\n",
      "[[  5.71630243e-05]]\n",
      "[[ 0.00535541]]\n",
      "[[ 0.00048592]]\n",
      "[[ 0.00096058]]\n",
      "[[ 0.00080677]]\n",
      "[[  1.54387478e-06]]\n",
      "[[  4.40274401e-05]]\n",
      "[[ 0.00020876]]\n",
      "[[  7.97192986e-05]]\n",
      "[[  3.03188488e-05]]\n",
      "[[ 0.00052638]]\n",
      "[[ 0.00016816]]\n",
      "[[  4.40262421e-05]]\n",
      "[[  6.63138865e-05]]\n",
      "[[ 0.00037466]]\n",
      "[[ 0.00284485]]\n",
      "[[ 0.00013348]]\n",
      "[[ 0.00045128]]\n",
      "[[ 0.00026848]]\n",
      "[[ 0.00026975]]\n",
      "[[ 0.01829491]]\n",
      "[[ 0.00031582]]\n",
      "-------results for values of 3--------\n",
      "[[  6.67222478e-07]]\n",
      "[[  8.80703885e-08]]\n",
      "[[  2.37477592e-07]]\n",
      "[[  6.42293963e-08]]\n",
      "[[  5.43681921e-09]]\n",
      "[[  3.93019415e-06]]\n",
      "[[  2.34482745e-08]]\n",
      "[[  4.06887647e-08]]\n",
      "[[  9.84007379e-08]]\n",
      "[[  4.70362348e-06]]\n",
      "[[  4.62586971e-07]]\n",
      "[[  4.26713396e-07]]\n",
      "[[  4.25017804e-07]]\n",
      "[[  3.05061257e-07]]\n",
      "[[  8.81133772e-07]]\n",
      "[[  6.17312061e-07]]\n",
      "[[  3.39569926e-08]]\n",
      "[[  2.86014776e-10]]\n",
      "[[  6.72874285e-07]]\n",
      "[[  2.19444606e-07]]\n",
      "[[  1.52632068e-06]]\n",
      "[[  2.54522965e-09]]\n",
      "[[  3.35999607e-06]]\n",
      "[[  3.79846047e-07]]\n",
      "[[  2.13400373e-07]]\n",
      "[[  6.58924608e-06]]\n",
      "[[  6.90015940e-06]]\n",
      "[[  1.00529715e-07]]\n",
      "[[  4.55828179e-07]]\n",
      "[[  3.88266144e-08]]\n",
      "[[  2.55145497e-07]]\n",
      "[[  8.09177178e-08]]\n",
      "[[  3.44045914e-06]]\n",
      "[[  1.89201914e-07]]\n",
      "[[  3.95595103e-08]]\n",
      "[[  2.93068796e-07]]\n",
      "[[  7.60467134e-07]]\n",
      "[[  9.99967176e-06]]\n",
      "[[  1.00977281e-06]]\n",
      "[[  1.72201539e-07]]\n",
      "[[  2.63930226e-06]]\n",
      "[[  6.67222478e-07]]\n",
      "[[  7.88288346e-08]]\n",
      "[[  1.59442603e-07]]\n",
      "[[  1.01533076e-06]]\n",
      "[[  9.27738539e-07]]\n",
      "[[  1.61350567e-06]]\n",
      "[[  8.39983496e-07]]\n",
      "[[  2.25645477e-06]]\n"
     ]
    }
   ],
   "source": [
    "test(w_1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It can be seen that the iris of type 1 can be distinguished by pr > 0.9"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classify 2 versus 1,3\n",
    "### In this case 1 and 3 will be interpreted as 0 and 2 as 1 because logistic regression is binary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.8207203 ]\n",
      " [-1.87483639]\n",
      " [ 0.60606818]\n",
      " [-1.87280186]]\n"
     ]
    }
   ],
   "source": [
    "y = np.matrix(df_iris.c[:]).T\n",
    "y[y == 1] = 0\n",
    "y[y == 2] = 1\n",
    "y[y == 3] = 0\n",
    "w_2 = w_calculator(FeatureVector,y,0.0001)\n",
    "print(w_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test w_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------results for values of 1--------\n",
      "[[ 0.12984626]]\n",
      "[[ 0.24433745]]\n",
      "[[ 0.1507459]]\n",
      "[[ 0.18210693]]\n",
      "[[ 0.10230476]]\n",
      "[[ 0.06921716]]\n",
      "[[ 0.09009501]]\n",
      "[[ 0.14978347]]\n",
      "[[ 0.20555647]]\n",
      "[[ 0.25566156]]\n",
      "[[ 0.12233915]]\n",
      "[[ 0.1370705]]\n",
      "[[ 0.26428078]]\n",
      "[[ 0.1657549]]\n",
      "[[ 0.08421208]]\n",
      "[[ 0.03194748]]\n",
      "[[ 0.05513776]]\n",
      "[[ 0.11011187]]\n",
      "[[ 0.12155385]]\n",
      "[[ 0.06969117]]\n",
      "[[ 0.21639518]]\n",
      "[[ 0.06970436]]\n",
      "[[ 0.06050652]]\n",
      "[[ 0.12927643]]\n",
      "[[ 0.16002835]]\n",
      "[[ 0.28378592]]\n",
      "[[ 0.11402673]]\n",
      "[[ 0.14683544]]\n",
      "[[ 0.1634522]]\n",
      "[[ 0.17552833]]\n",
      "[[ 0.2179954]]\n",
      "[[ 0.14398576]]\n",
      "[[ 0.06313514]]\n",
      "[[ 0.05282916]]\n",
      "[[ 0.25566156]]\n",
      "[[ 0.17607731]]\n",
      "[[ 0.16319606]]\n",
      "[[ 0.25566156]]\n",
      "[[ 0.16797999]]\n",
      "[[ 0.1605383]]\n",
      "[[ 0.09688915]]\n",
      "[[ 0.40303519]]\n",
      "[[ 0.12185511]]\n",
      "[[ 0.06835113]]\n",
      "[[ 0.07335255]]\n",
      "[[ 0.19807063]]\n",
      "[[ 0.08757934]]\n",
      "[[ 0.14801844]]\n",
      "[[ 0.11379647]]\n",
      "[[ 0.16666867]]\n",
      "-------results for values of 2--------\n",
      "[[ 0.30390679]]\n",
      "[[ 0.5028771]]\n",
      "[[ 0.54772136]]\n",
      "[[ 0.5159573]]\n",
      "[[ 0.43075491]]\n",
      "[[ 0.23787147]]\n",
      "[[ 0.41317516]]\n",
      "[[ 0.58248906]]\n",
      "[[ 0.25873201]]\n",
      "[[ 0.6461973]]\n",
      "[[ 0.2599986]]\n",
      "[[ 0.794318]]\n",
      "[[ 0.44916775]]\n",
      "[[ 0.25090549]]\n",
      "[[ 0.43330253]]\n",
      "[[ 0.24780164]]\n",
      "[[ 0.57693338]]\n",
      "[[ 0.70722196]]\n",
      "[[ 0.55293255]]\n",
      "[[ 0.16532313]]\n",
      "[[ 0.43695784]]\n",
      "[[ 0.65565218]]\n",
      "[[ 0.58855866]]\n",
      "[[ 0.49675883]]\n",
      "[[ 0.45934549]]\n",
      "[[ 0.64989115]]\n",
      "[[ 0.4306731]]\n",
      "[[ 0.35558016]]\n",
      "[[ 0.51298904]]\n",
      "[[ 0.56398016]]\n",
      "[[ 0.59483862]]\n",
      "[[ 0.45373848]]\n",
      "[[ 0.48918376]]\n",
      "[[ 0.21848525]]\n",
      "[[ 0.15196159]]\n",
      "[[ 0.43196238]]\n",
      "[[ 0.74847054]]\n",
      "[[ 0.27324147]]\n",
      "[[ 0.45425353]]\n",
      "[[ 0.51467698]]\n",
      "[[ 0.38885581]]\n",
      "[[ 0.51562331]]\n",
      "[[ 0.47968642]]\n",
      "[[ 0.41212669]]\n",
      "[[ 0.34337739]]\n",
      "[[ 0.34342326]]\n",
      "[[ 0.45583814]]\n",
      "[[ 0.32228405]]\n",
      "[[ 0.37257189]]\n",
      "-------results for values of 3--------\n",
      "[[ 0.31664]]\n",
      "[[ 0.46145744]]\n",
      "[[ 0.43941189]]\n",
      "[[ 0.29012112]]\n",
      "[[ 0.66376797]]\n",
      "[[ 0.24563718]]\n",
      "[[ 0.73133745]]\n",
      "[[ 0.72229758]]\n",
      "[[ 0.13880946]]\n",
      "[[ 0.21090973]]\n",
      "[[ 0.46117568]]\n",
      "[[ 0.34454159]]\n",
      "[[ 0.32645821]]\n",
      "[[ 0.13088676]]\n",
      "[[ 0.13679728]]\n",
      "[[ 0.41885011]]\n",
      "[[ 0.29643354]]\n",
      "[[ 0.78908151]]\n",
      "[[ 0.73513226]]\n",
      "[[ 0.23337305]]\n",
      "[[ 0.19319564]]\n",
      "[[ 0.79980203]]\n",
      "[[ 0.42731126]]\n",
      "[[ 0.237501]]\n",
      "[[ 0.5436574]]\n",
      "[[ 0.34910124]]\n",
      "[[ 0.26513949]]\n",
      "[[ 0.36915721]]\n",
      "[[ 0.69069905]]\n",
      "[[ 0.7236182]]\n",
      "[[ 0.37579025]]\n",
      "[[ 0.32670757]]\n",
      "[[ 0.55051232]]\n",
      "[[ 0.71174586]]\n",
      "[[ 0.52113958]]\n",
      "[[ 0.0907387]]\n",
      "[[ 0.35501861]]\n",
      "[[ 0.23828528]]\n",
      "[[ 0.30807411]]\n",
      "[[ 0.19561816]]\n",
      "[[ 0.20334473]]\n",
      "[[ 0.31664]]\n",
      "[[ 0.24044882]]\n",
      "[[ 0.12835861]]\n",
      "[[ 0.2172826]]\n",
      "[[ 0.48886791]]\n",
      "[[ 0.29237513]]\n",
      "[[ 0.08942679]]\n",
      "[[ 0.25685966]]\n"
     ]
    }
   ],
   "source": [
    "test(w_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It can be seen that the iris of type 2 is much harder to distinguish"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classify 3 versus 1,2\n",
    "### In this case 1 and 2 will be interpreted as 0 and 3 as 1 because logistic regression is binary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-4.81285544]\n",
      " [-4.963746  ]\n",
      " [ 6.37297801]\n",
      " [ 7.85661824]]\n"
     ]
    }
   ],
   "source": [
    "y = np.matrix(df_iris.c[:]).T\n",
    "y[y < 3] = 0\n",
    "y[y == 3] = 1\n",
    "w_3 = w_calculator(FeatureVector,y,0.001)\n",
    "print(w_3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test w_3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------results for values of 1--------\n",
      "[[  2.25018043e-14]]\n",
      "[[  7.04888467e-13]]\n",
      "[[  3.61609057e-13]]\n",
      "[[  3.43859908e-12]]\n",
      "[[  2.21648221e-14]]\n",
      "[[  2.37470312e-14]]\n",
      "[[  8.99698444e-13]]\n",
      "[[  1.13131793e-13]]\n",
      "[[  1.28466957e-11]]\n",
      "[[  3.69925253e-13]]\n",
      "[[  3.72209792e-15]]\n",
      "[[  5.60271975e-13]]\n",
      "[[  5.19913945e-13]]\n",
      "[[  8.52520815e-13]]\n",
      "[[  1.80994828e-17]]\n",
      "[[  1.30959427e-16]]\n",
      "[[  1.85570803e-15]]\n",
      "[[  4.93657727e-14]]\n",
      "[[  4.19676362e-15]]\n",
      "[[  2.10611756e-14]]\n",
      "[[  5.90277027e-14]]\n",
      "[[  7.59038759e-14]]\n",
      "[[  1.18752620e-14]]\n",
      "[[  4.33823810e-12]]\n",
      "[[  3.79073700e-12]]\n",
      "[[  1.55829826e-12]]\n",
      "[[  1.02985665e-12]]\n",
      "[[  2.63010209e-14]]\n",
      "[[  2.28439099e-14]]\n",
      "[[  2.44660610e-12]]\n",
      "[[  2.48380301e-12]]\n",
      "[[  7.94187841e-14]]\n",
      "[[  6.09996739e-16]]\n",
      "[[  1.01655170e-16]]\n",
      "[[  3.69925253e-13]]\n",
      "[[  4.51237737e-14]]\n",
      "[[  1.73525698e-15]]\n",
      "[[  3.69925253e-13]]\n",
      "[[  4.13469775e-12]]\n",
      "[[  6.99141388e-14]]\n",
      "[[  4.22348228e-14]]\n",
      "[[  1.80984738e-10]]\n",
      "[[  1.53213935e-12]]\n",
      "[[  3.01731893e-12]]\n",
      "[[  5.91276852e-13]]\n",
      "[[  2.50235347e-12]]\n",
      "[[  1.81572023e-14]]\n",
      "[[  1.10671010e-12]]\n",
      "[[  6.02292494e-15]]\n",
      "[[  9.82612987e-14]]\n",
      "-------results for values of 2--------\n",
      "[[ 0.00197949]]\n",
      "[[ 0.0037442]]\n",
      "[[ 0.10139635]]\n",
      "[[ 0.01660311]]\n",
      "[[ 0.08019061]]\n",
      "[[ 0.01510119]]\n",
      "[[ 0.00134699]]\n",
      "[[ 0.00131789]]\n",
      "[[ 0.07075937]]\n",
      "[[ 0.02125398]]\n",
      "[[ 0.00870038]]\n",
      "[[ 0.00157985]]\n",
      "[[ 0.05726911]]\n",
      "[[ 0.00027719]]\n",
      "[[ 0.0001853]]\n",
      "[[ 0.20102505]]\n",
      "[[ 0.00065459]]\n",
      "[[ 0.42638481]]\n",
      "[[ 0.00283052]]\n",
      "[[ 0.61120328]]\n",
      "[[ 0.0005251]]\n",
      "[[ 0.570077]]\n",
      "[[ 0.02031307]]\n",
      "[[ 0.00051044]]\n",
      "[[ 0.00049241]]\n",
      "[[ 0.0064555]]\n",
      "[[ 0.12828274]]\n",
      "[[ 0.05685759]]\n",
      "[[  3.80346727e-05]]\n",
      "[[ 0.00397361]]\n",
      "[[ 0.00096054]]\n",
      "[[ 0.00088052]]\n",
      "[[ 0.94232521]]\n",
      "[[ 0.39715608]]\n",
      "[[ 0.01093405]]\n",
      "[[ 0.00274337]]\n",
      "[[ 0.02980277]]\n",
      "[[ 0.00406846]]\n",
      "[[ 0.0401346]]\n",
      "[[ 0.1292733]]\n",
      "[[ 0.01917678]]\n",
      "[[ 0.00273074]]\n",
      "[[ 0.00136744]]\n",
      "[[ 0.0331182]]\n",
      "[[ 0.00217171]]\n",
      "[[ 0.00778278]]\n",
      "[[ 0.00133544]]\n",
      "[[ 0.00010167]]\n",
      "[[ 0.00676671]]\n",
      "-------results for values of 3--------\n",
      "[[ 0.99779118]]\n",
      "[[ 0.99354841]]\n",
      "[[ 0.99402832]]\n",
      "[[ 0.99968826]]\n",
      "[[ 0.99916864]]\n",
      "[[ 0.99762953]]\n",
      "[[ 0.99153208]]\n",
      "[[ 0.99842148]]\n",
      "[[ 0.99751419]]\n",
      "[[ 0.74036697]]\n",
      "[[ 0.98901272]]\n",
      "[[ 0.9807654]]\n",
      "[[ 0.99956315]]\n",
      "[[ 0.99992845]]\n",
      "[[ 0.9942953]]\n",
      "[[ 0.95340264]]\n",
      "[[ 0.98308161]]\n",
      "[[ 0.99999432]]\n",
      "[[ 0.97921345]]\n",
      "[[ 0.9950505]]\n",
      "[[ 0.9977405]]\n",
      "[[ 0.99942163]]\n",
      "[[ 0.83840412]]\n",
      "[[ 0.98520237]]\n",
      "[[ 0.8633308]]\n",
      "[[ 0.72988018]]\n",
      "[[ 0.75396069]]\n",
      "[[ 0.99943989]]\n",
      "[[ 0.49751614]]\n",
      "[[ 0.98646864]]\n",
      "[[ 0.40528373]]\n",
      "[[ 0.99974461]]\n",
      "[[ 0.51690284]]\n",
      "[[ 0.98815325]]\n",
      "[[ 0.99327492]]\n",
      "[[ 0.99935579]]\n",
      "[[ 0.95272769]]\n",
      "[[ 0.72388931]]\n",
      "[[ 0.91024781]]\n",
      "[[ 0.99900405]]\n",
      "[[ 0.87826424]]\n",
      "[[ 0.99779118]]\n",
      "[[ 0.99914143]]\n",
      "[[ 0.99935204]]\n",
      "[[ 0.98324791]]\n",
      "[[ 0.98307861]]\n",
      "[[ 0.93571134]]\n",
      "[[ 0.99688331]]\n",
      "[[ 0.96633382]]\n"
     ]
    }
   ],
   "source": [
    "test(w_3)"
   ]
  }
 ],
 "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}