hybridLinUCB.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "from movielens import MovieLens\n",
    "from hybridLinUCB import HybridLinUCB\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from matplotlib.pylab import rcParams\n",
    "rcParams['figure.figsize'] = 7, 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data\\movielens\\ml-100k.zip is already downloaded.\n",
      "Created a rating matrix of shape=(943, 1682) and dtype=float64 from u.data.\n",
      "Positive ratings (>=4) turned to 1, negative to -1, unknown to 0\n",
      "Created a genre matrix of shape=(1682, 19) and dtype=float64 from u.item.\n",
      "Created a titles matrix of shape=(1682,) and dtype=object from u.item.\n",
      "Statistics about self.R:\n",
      "Number of users:           943\n",
      "Number of items:           1682\n",
      "Total number of ratings:   1586126\n",
      "Known ratings:             100000\n",
      "Known positive ratings:    55375\n",
      "Known negative ratings:    44625\n",
      "Ratio of known ratings:    0.06304669364224531\n",
      "Ratio of positive ratings: 0.034912106604393346\n",
      "Ratio of negative ratings: 0.028134587037851974\n",
      "Avg number of positive ratings per user: 58.72216330858961 +- 54.67211827074995\n",
      "Avg number of negative ratings per user: 47.32237539766702 +- 58.41143787053165\n",
      "Shrinked rating matrix from (943, 1682) to (56, 100).\n",
      "\n",
      "After shrinking:\n",
      "Number of users:           56\n",
      "Number of items:           100\n",
      "Total number of ratings:   5600\n",
      "Known ratings:             903\n",
      "Known positive ratings:    606\n",
      "Known negative ratings:    297\n",
      "Ratio of known ratings:    0.16125\n",
      "Ratio of positive ratings: 0.10821428571428572\n",
      "Ratio of negative ratings: 0.05303571428571428\n",
      "Avg number of positive ratings per user: 10.821428571428571 +- 11.918034268130233\n",
      "Avg number of negative ratings per user: 5.303571428571429 +- 8.209140469312118\n",
      "\n",
      "Initializing matrix A of shape (100, 119, 119) which will require 11.3288MB of memory.\n",
      "\n",
      "LinUCB successfully initialized.\n"
     ]
    }
   ],
   "source": [
    "ucb = HybridLinUCB(alpha=0.1, dataset=None, max_items=100, allow_selecting_known_arms=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R:\n",
      " [[ 1. -1.  1. ...,  1. -1.  1.]\n",
      " [ 1.  0.  0. ...,  0.  0.  1.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " ..., \n",
      " [ 1.  0.  0. ...,  0.  0.  1.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " [ 1.  0.  0. ...,  1.  0.  0.]]\n"
     ]
    }
   ],
   "source": [
    "print('R:\\n', ucb.dataset.R)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished epoch 0/50 with avg reward 0.17857142857142858 in 1.750478744506836s\n",
      "Finished epoch 1/50 with avg reward 0.25 in 1.7018945217132568s\n",
      "Finished epoch 2/50 with avg reward 0.39285714285714285 in 1.695727825164795s\n",
      "Finished epoch 3/50 with avg reward 0.4107142857142857 in 1.6814639568328857s\n",
      "Finished epoch 4/50 with avg reward 0.5178571428571429 in 1.6740617752075195s\n",
      "Finished epoch 5/50 with avg reward 0.5535714285714286 in 1.6604721546173096s\n",
      "Finished epoch 6/50 with avg reward 0.5357142857142857 in 1.7458255290985107s\n",
      "Finished epoch 7/50 with avg reward 0.6964285714285714 in 1.8287920951843262s\n",
      "Finished epoch 8/50 with avg reward 0.6607142857142857 in 1.714789867401123s\n",
      "Finished epoch 9/50 with avg reward 0.6785714285714286 in 1.7897109985351562s\n",
      "Finished epoch 10/50 with avg reward 0.6428571428571429 in 1.7908804416656494s\n",
      "Finished epoch 11/50 with avg reward 0.7678571428571429 in 1.6843452453613281s\n",
      "Finished epoch 12/50 with avg reward 0.7321428571428571 in 1.693253517150879s\n",
      "Finished epoch 13/50 with avg reward 0.7678571428571429 in 1.742851734161377s\n",
      "Finished epoch 14/50 with avg reward 0.7678571428571429 in 1.8221018314361572s\n",
      "Finished epoch 15/50 with avg reward 0.8214285714285714 in 1.8044462203979492s\n",
      "Finished epoch 16/50 with avg reward 0.8035714285714286 in 1.75416898727417s\n",
      "Finished epoch 17/50 with avg reward 0.8035714285714286 in 1.6956160068511963s\n",
      "Finished epoch 18/50 with avg reward 0.8392857142857143 in 1.7304728031158447s\n",
      "Finished epoch 19/50 with avg reward 0.8392857142857143 in 1.6806674003601074s\n",
      "Finished epoch 20/50 with avg reward 0.8214285714285714 in 1.748338222503662s\n",
      "Finished epoch 21/50 with avg reward 0.8392857142857143 in 1.839202880859375s\n",
      "Finished epoch 22/50 with avg reward 0.8571428571428571 in 1.7647366523742676s\n",
      "Finished epoch 23/50 with avg reward 0.8392857142857143 in 1.800455093383789s\n",
      "Finished epoch 24/50 with avg reward 0.8571428571428571 in 1.7805113792419434s\n",
      "Finished epoch 25/50 with avg reward 0.8214285714285714 in 1.7618687152862549s\n",
      "Finished epoch 26/50 with avg reward 0.8214285714285714 in 1.761312484741211s\n",
      "Finished epoch 27/50 with avg reward 0.8035714285714286 in 1.7098705768585205s\n",
      "Finished epoch 28/50 with avg reward 0.8214285714285714 in 1.7373664379119873s\n",
      "Finished epoch 29/50 with avg reward 0.8392857142857143 in 1.718550443649292s\n",
      "Finished epoch 30/50 with avg reward 0.8928571428571429 in 1.7102465629577637s\n",
      "Finished epoch 31/50 with avg reward 0.8571428571428571 in 1.7103793621063232s\n",
      "Finished epoch 32/50 with avg reward 0.8392857142857143 in 1.6831531524658203s\n",
      "Finished epoch 33/50 with avg reward 0.9285714285714286 in 1.7274296283721924s\n",
      "Finished epoch 34/50 with avg reward 0.875 in 1.7818787097930908s\n",
      "Finished epoch 35/50 with avg reward 0.875 in 1.6770353317260742s\n",
      "Finished epoch 36/50 with avg reward 0.8928571428571429 in 1.704763650894165s\n",
      "Finished epoch 37/50 with avg reward 0.9285714285714286 in 1.698009729385376s\n",
      "Finished epoch 38/50 with avg reward 0.9107142857142857 in 1.7477633953094482s\n",
      "Finished epoch 39/50 with avg reward 0.9107142857142857 in 1.7294504642486572s\n",
      "Finished epoch 40/50 with avg reward 0.9285714285714286 in 1.712935209274292s\n",
      "Finished epoch 41/50 with avg reward 0.8928571428571429 in 1.7294156551361084s\n",
      "Finished epoch 42/50 with avg reward 0.9107142857142857 in 1.776698112487793s\n",
      "Finished epoch 43/50 with avg reward 0.9285714285714286 in 1.7602694034576416s\n",
      "Finished epoch 44/50 with avg reward 0.9464285714285714 in 1.7506699562072754s\n",
      "Finished epoch 45/50 with avg reward 0.9464285714285714 in 1.694836139678955s\n",
      "Finished epoch 46/50 with avg reward 0.9107142857142857 in 1.7082233428955078s\n",
      "Finished epoch 47/50 with avg reward 0.8928571428571429 in 1.7098522186279297s\n",
      "Finished epoch 48/50 with avg reward 0.9285714285714286 in 1.7239134311676025s\n",
      "Finished epoch 49/50 with avg reward 0.9464285714285714 in 1.7143592834472656s\n"
     ]
    }
   ],
   "source": [
    "avg_reward = ucb.run(num_epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x260375eec50>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb0AAAFNCAYAAACUvLFdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkUgkKoY06SUSiUSiYkiTXiKRSCQqhj8B9WQYBQzl0fYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x260375bc400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(range(len(avg_reward)), avg_reward)\n",
    "plt.ylabel('average reward in epoch')\n",
    "plt.ylim(0.2, 1)\n",
    "plt.xlabel('epoch (1 epoch = recommended to all users)')\n",
    "plt.title('HybridLinUCB: Fixed rewards')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data\\movielens\\ml-100k.zip is already downloaded.\n",
      "Created a rating matrix of shape=(943, 1682) and dtype=float64 from u.data.\n",
      "Positive ratings (>=4) turned to 1, negative to -1, unknown to 0\n",
      "Created a genre matrix of shape=(1682, 19) and dtype=float64 from u.item.\n",
      "Created a titles matrix of shape=(1682,) and dtype=object from u.item.\n",
      "Statistics about self.R:\n",
      "Number of users:           943\n",
      "Number of items:           1682\n",
      "Total number of ratings:   1586126\n",
      "Known ratings:             100000\n",
      "Known positive ratings:    55375\n",
      "Known negative ratings:    44625\n",
      "Ratio of known ratings:    0.06304669364224531\n",
      "Ratio of positive ratings: 0.034912106604393346\n",
      "Ratio of negative ratings: 0.028134587037851974\n",
      "Avg number of positive ratings per user: 58.72216330858961 +- 54.67211827074995\n",
      "Avg number of negative ratings per user: 47.32237539766702 +- 58.41143787053165\n",
      "Shrinked rating matrix from (943, 1682) to (56, 100).\n",
      "\n",
      "After shrinking:\n",
      "Number of users:           56\n",
      "Number of items:           100\n",
      "Total number of ratings:   5600\n",
      "Known ratings:             903\n",
      "Known positive ratings:    606\n",
      "Known negative ratings:    297\n",
      "Ratio of known ratings:    0.16125\n",
      "Ratio of positive ratings: 0.10821428571428572\n",
      "Ratio of negative ratings: 0.05303571428571428\n",
      "Avg number of positive ratings per user: 10.821428571428571 +- 11.918034268130233\n",
      "Avg number of negative ratings per user: 5.303571428571429 +- 8.209140469312118\n",
      "\n",
      "Initializing matrix A of shape (100, 119, 119) which will require 11.3288MB of memory.\n",
      "\n",
      "LinUCB successfully initialized.\n"
     ]
    }
   ],
   "source": [
    "ucb_stochastic = HybridLinUCB(alpha=0.1, dataset=None, max_items=100, \n",
    "                        allow_selecting_known_arms=True, fixed_rewards=False, prob_reward_p=0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R:\n",
      " [[ 1. -1.  1. ...,  1. -1.  1.]\n",
      " [ 1.  0.  0. ...,  0.  0.  1.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " ..., \n",
      " [ 1.  0.  0. ...,  0.  0.  1.]\n",
      " [ 0.  0.  0. ...,  0.  0.  0.]\n",
      " [ 1.  0.  0. ...,  1.  0.  0.]]\n"
     ]
    }
   ],
   "source": [
    "print('R:\\n', ucb_stochastic.dataset.R)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished epoch 0/50 with avg reward 0.30357142857142855 in 1.748520851135254s\n",
      "Finished epoch 1/50 with avg reward 0.5535714285714286 in 1.7934753894805908s\n",
      "Finished epoch 2/50 with avg reward 0.48214285714285715 in 1.7535381317138672s\n",
      "Finished epoch 3/50 with avg reward 0.5714285714285714 in 1.7890560626983643s\n",
      "Finished epoch 4/50 with avg reward 0.5714285714285714 in 1.8073928356170654s\n",
      "Finished epoch 5/50 with avg reward 0.5714285714285714 in 1.812791347503662s\n",
      "Finished epoch 6/50 with avg reward 0.5714285714285714 in 1.7791979312896729s\n",
      "Finished epoch 7/50 with avg reward 0.6428571428571429 in 1.7468101978302002s\n",
      "Finished epoch 8/50 with avg reward 0.6071428571428571 in 1.6790573596954346s\n",
      "Finished epoch 9/50 with avg reward 0.625 in 1.746351957321167s\n",
      "Finished epoch 10/50 with avg reward 0.6428571428571429 in 1.7344410419464111s\n",
      "Finished epoch 11/50 with avg reward 0.6071428571428571 in 1.843202829360962s\n",
      "Finished epoch 12/50 with avg reward 0.6785714285714286 in 1.8209164142608643s\n",
      "Finished epoch 13/50 with avg reward 0.6428571428571429 in 1.8320651054382324s\n",
      "Finished epoch 14/50 with avg reward 0.6607142857142857 in 1.8215398788452148s\n",
      "Finished epoch 15/50 with avg reward 0.5892857142857143 in 1.729494333267212s\n",
      "Finished epoch 16/50 with avg reward 0.6607142857142857 in 1.7352945804595947s\n",
      "Finished epoch 17/50 with avg reward 0.6785714285714286 in 1.7602496147155762s\n",
      "Finished epoch 18/50 with avg reward 0.6607142857142857 in 1.7683212757110596s\n",
      "Finished epoch 19/50 with avg reward 0.6964285714285714 in 1.697112798690796s\n",
      "Finished epoch 20/50 with avg reward 0.6785714285714286 in 1.6891827583312988s\n",
      "Finished epoch 21/50 with avg reward 0.6785714285714286 in 1.7493321895599365s\n",
      "Finished epoch 22/50 with avg reward 0.6785714285714286 in 1.7260868549346924s\n",
      "Finished epoch 23/50 with avg reward 0.7142857142857143 in 1.756211280822754s\n",
      "Finished epoch 24/50 with avg reward 0.7142857142857143 in 1.7074334621429443s\n",
      "Finished epoch 25/50 with avg reward 0.6785714285714286 in 1.8026928901672363s\n",
      "Finished epoch 26/50 with avg reward 0.6964285714285714 in 1.78187894821167s\n",
      "Finished epoch 27/50 with avg reward 0.6607142857142857 in 1.725114107131958s\n",
      "Finished epoch 28/50 with avg reward 0.7678571428571429 in 1.7646830081939697s\n",
      "Finished epoch 29/50 with avg reward 0.8035714285714286 in 1.7060692310333252s\n",
      "Finished epoch 30/50 with avg reward 0.8214285714285714 in 1.7053234577178955s\n",
      "Finished epoch 31/50 with avg reward 0.7678571428571429 in 1.7799861431121826s\n",
      "Finished epoch 32/50 with avg reward 0.7321428571428571 in 1.7605571746826172s\n",
      "Finished epoch 33/50 with avg reward 0.6964285714285714 in 1.7521774768829346s\n",
      "Finished epoch 34/50 with avg reward 0.8035714285714286 in 1.7175846099853516s\n",
      "Finished epoch 35/50 with avg reward 0.6785714285714286 in 1.814420223236084s\n",
      "Finished epoch 36/50 with avg reward 0.75 in 1.8053712844848633s\n",
      "Finished epoch 37/50 with avg reward 0.7142857142857143 in 1.7628092765808105s\n",
      "Finished epoch 38/50 with avg reward 0.8035714285714286 in 1.754706621170044s\n",
      "Finished epoch 39/50 with avg reward 0.7142857142857143 in 1.726529836654663s\n",
      "Finished epoch 40/50 with avg reward 0.7321428571428571 in 1.7202081680297852s\n",
      "Finished epoch 41/50 with avg reward 0.7857142857142857 in 1.8135907649993896s\n",
      "Finished epoch 42/50 with avg reward 0.7678571428571429 in 1.7712664604187012s\n",
      "Finished epoch 43/50 with avg reward 0.75 in 1.7777793407440186s\n",
      "Finished epoch 44/50 with avg reward 0.7142857142857143 in 1.7925455570220947s\n",
      "Finished epoch 45/50 with avg reward 0.7857142857142857 in 1.7052061557769775s\n",
      "Finished epoch 46/50 with avg reward 0.8214285714285714 in 1.711728811264038s\n",
      "Finished epoch 47/50 with avg reward 0.7857142857142857 in 1.703730583190918s\n",
      "Finished epoch 48/50 with avg reward 0.8035714285714286 in 1.7389450073242188s\n",
      "Finished epoch 49/50 with avg reward 0.7142857142857143 in 1.7418231964111328s\n"
     ]
    }
   ],
   "source": [
    "avg_reward_stochastic = ucb_stochastic.run(num_epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x260376843c8>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAb0AAAFNCAYAAACUvLFdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmZlZbbjomZlZbfw/LyXJj8uBCQ8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26037636780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(range(len(avg_reward_stochastic)), avg_reward_stochastic)\n",
    "plt.ylabel('average reward in epoch')\n",
    "plt.ylim(0.2, 1)\n",
    "plt.xlabel('epoch (1 epoch = recommended to all users)')\n",
    "plt.title('HybridLinUCB: Stochastic rewards with p=0.9')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data\\movielens\\ml-100k.zip is already downloaded.\n",
      "Created a rating matrix of shape=(943, 1682) and dtype=float64 from u.data.\n",
      "Positive ratings (>=4) turned to 1, negative to -1, unknown to 0\n",
      "Created a genre matrix of shape=(1682, 19) and dtype=float64 from u.item.\n",
      "Created a titles matrix of shape=(1682,) and dtype=object from u.item.\n",
      "Statistics about self.R:\n",
      "Number of users:           943\n",
      "Number of items:           1682\n",
      "Total number of ratings:   1586126\n",
      "Known ratings:             100000\n",
      "Known positive ratings:    55375\n",
      "Known negative ratings:    44625\n",
      "Ratio of known ratings:    0.06304669364224531\n",
      "Ratio of positive ratings: 0.034912106604393346\n",
      "Ratio of negative ratings: 0.028134587037851974\n",
      "Avg number of positive ratings per user: 58.72216330858961 +- 54.67211827074995\n",
      "Avg number of negative ratings per user: 47.32237539766702 +- 58.41143787053165\n",
      "Shrinked rating matrix from (943, 1682) to (56, 100).\n",
      "\n",
      "After shrinking:\n",
      "Number of users:           56\n",
      "Number of items:           100\n",
      "Total number of ratings:   5600\n",
      "Known ratings:             903\n",
      "Known positive ratings:    606\n",
      "Known negative ratings:    297\n",
      "Ratio of known ratings:    0.16125\n",
      "Ratio of positive ratings: 0.10821428571428572\n",
      "Ratio of negative ratings: 0.05303571428571428\n",
      "Avg number of positive ratings per user: 10.821428571428571 +- 11.918034268130233\n",
      "Avg number of negative ratings per user: 5.303571428571429 +- 8.209140469312118\n",
      "\n",
      "Initializing matrix A of shape (100, 119, 119) which will require 11.3288MB of memory.\n",
      "\n",
      "LinUCB successfully initialized.\n"
     ]
    }
   ],
   "source": [
    "ucb_stochastic_100 = HybridLinUCB(alpha=0.1, dataset=None, max_items=100, \n",
    "                        allow_selecting_known_arms=True, fixed_rewards=False, prob_reward_p=0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Finished epoch 0/100 with avg reward 0.30357142857142855 in 2.194092273712158s\n",
      "Finished epoch 1/100 with avg reward 0.5535714285714286 in 2.196359395980835s\n",
      "Finished epoch 2/100 with avg reward 0.48214285714285715 in 2.1993517875671387s\n",
      "Finished epoch 3/100 with avg reward 0.5714285714285714 in 2.1957151889801025s\n",
      "Finished epoch 4/100 with avg reward 0.5714285714285714 in 2.1858274936676025s\n",
      "Finished epoch 5/100 with avg reward 0.5714285714285714 in 2.1821651458740234s\n",
      "Finished epoch 6/100 with avg reward 0.5714285714285714 in 2.165241241455078s\n",
      "Finished epoch 7/100 with avg reward 0.6428571428571429 in 2.193681478500366s\n",
      "Finished epoch 8/100 with avg reward 0.6071428571428571 in 2.1821019649505615s\n",
      "Finished epoch 9/100 with avg reward 0.625 in 2.1016576290130615s\n",
      "Finished epoch 10/100 with avg reward 0.6428571428571429 in 2.140442132949829s\n",
      "Finished epoch 11/100 with avg reward 0.6071428571428571 in 2.1395950317382812s\n",
      "Finished epoch 12/100 with avg reward 0.6785714285714286 in 2.14200496673584s\n",
      "Finished epoch 13/100 with avg reward 0.6428571428571429 in 2.0825142860412598s\n",
      "Finished epoch 14/100 with avg reward 0.6607142857142857 in 2.114504098892212s\n",
      "Finished epoch 15/100 with avg reward 0.5892857142857143 in 2.1718075275421143s\n",
      "Finished epoch 16/100 with avg reward 0.6607142857142857 in 2.1072230339050293s\n",
      "Finished epoch 17/100 with avg reward 0.6785714285714286 in 2.1290102005004883s\n",
      "Finished epoch 18/100 with avg reward 0.6607142857142857 in 2.1607754230499268s\n",
      "Finished epoch 19/100 with avg reward 0.6964285714285714 in 2.167294502258301s\n",
      "Finished epoch 20/100 with avg reward 0.6785714285714286 in 2.0770530700683594s\n",
      "Finished epoch 21/100 with avg reward 0.6785714285714286 in 2.11110520362854s\n",
      "Finished epoch 22/100 with avg reward 0.6785714285714286 in 2.040354013442993s\n",
      "Finished epoch 23/100 with avg reward 0.7142857142857143 in 2.1633074283599854s\n",
      "Finished epoch 24/100 with avg reward 0.7142857142857143 in 2.1905245780944824s\n",
      "Finished epoch 25/100 with avg reward 0.6785714285714286 in 2.1556620597839355s\n",
      "Finished epoch 26/100 with avg reward 0.6964285714285714 in 2.0907208919525146s\n",
      "Finished epoch 27/100 with avg reward 0.6607142857142857 in 2.1917524337768555s\n",
      "Finished epoch 28/100 with avg reward 0.7678571428571429 in 2.1074018478393555s\n",
      "Finished epoch 29/100 with avg reward 0.8035714285714286 in 2.127044916152954s\n",
      "Finished epoch 30/100 with avg reward 0.8214285714285714 in 2.132843017578125s\n",
      "Finished epoch 31/100 with avg reward 0.7678571428571429 in 2.1289730072021484s\n",
      "Finished epoch 32/100 with avg reward 0.7321428571428571 in 2.0687766075134277s\n",
      "Finished epoch 33/100 with avg reward 0.6964285714285714 in 2.1707239151000977s\n",
      "Finished epoch 34/100 with avg reward 0.8035714285714286 in 2.1654961109161377s\n",
      "Finished epoch 35/100 with avg reward 0.6785714285714286 in 2.1334996223449707s\n",
      "Finished epoch 36/100 with avg reward 0.75 in 2.117527484893799s\n",
      "Finished epoch 37/100 with avg reward 0.7142857142857143 in 2.1586074829101562s\n",
      "Finished epoch 38/100 with avg reward 0.8035714285714286 in 2.1625494956970215s\n",
      "Finished epoch 39/100 with avg reward 0.7142857142857143 in 2.16343355178833s\n",
      "Finished epoch 40/100 with avg reward 0.7321428571428571 in 2.1276447772979736s\n",
      "Finished epoch 41/100 with avg reward 0.7857142857142857 in 2.125063180923462s\n",
      "Finished epoch 42/100 with avg reward 0.7678571428571429 in 2.1398653984069824s\n",
      "Finished epoch 43/100 with avg reward 0.75 in 2.143637180328369s\n",
      "Finished epoch 44/100 with avg reward 0.7142857142857143 in 2.0914266109466553s\n",
      "Finished epoch 45/100 with avg reward 0.7857142857142857 in 2.1763362884521484s\n",
      "Finished epoch 46/100 with avg reward 0.8214285714285714 in 2.190397262573242s\n",
      "Finished epoch 47/100 with avg reward 0.7857142857142857 in 2.207932949066162s\n",
      "Finished epoch 48/100 with avg reward 0.8035714285714286 in 2.180269956588745s\n",
      "Finished epoch 49/100 with avg reward 0.7142857142857143 in 2.1632659435272217s\n",
      "Finished epoch 50/100 with avg reward 0.7678571428571429 in 2.164156436920166s\n",
      "Finished epoch 51/100 with avg reward 0.7857142857142857 in 2.1201491355895996s\n",
      "Finished epoch 52/100 with avg reward 0.7678571428571429 in 2.186532974243164s\n",
      "Finished epoch 53/100 with avg reward 0.7857142857142857 in 2.1379973888397217s\n",
      "Finished epoch 54/100 with avg reward 0.7857142857142857 in 2.1604127883911133s\n",
      "Finished epoch 55/100 with avg reward 0.8214285714285714 in 2.077739953994751s\n",
      "Finished epoch 56/100 with avg reward 0.8928571428571429 in 2.1640055179595947s\n",
      "Finished epoch 57/100 with avg reward 0.875 in 2.1927361488342285s\n",
      "Finished epoch 58/100 with avg reward 0.8928571428571429 in 2.1336283683776855s\n",
      "Finished epoch 59/100 with avg reward 0.8392857142857143 in 2.1473701000213623s\n",
      "Finished epoch 60/100 with avg reward 0.8392857142857143 in 2.135578155517578s\n",
      "Finished epoch 61/100 with avg reward 0.8214285714285714 in 2.1821441650390625s\n",
      "Finished epoch 62/100 with avg reward 0.8928571428571429 in 2.186783790588379s\n",
      "Finished epoch 63/100 with avg reward 0.7142857142857143 in 2.1511127948760986s\n",
      "Finished epoch 64/100 with avg reward 0.8035714285714286 in 2.178368091583252s\n",
      "Finished epoch 65/100 with avg reward 0.8571428571428571 in 2.108289957046509s\n",
      "Finished epoch 66/100 with avg reward 0.6785714285714286 in 2.102919340133667s\n",
      "Finished epoch 67/100 with avg reward 0.7678571428571429 in 2.16042423248291s\n",
      "Finished epoch 68/100 with avg reward 0.8571428571428571 in 2.190430164337158s\n",
      "Finished epoch 69/100 with avg reward 0.8392857142857143 in 2.177781581878662s\n",
      "Finished epoch 70/100 with avg reward 0.8214285714285714 in 2.1046650409698486s\n",
      "Finished epoch 71/100 with avg reward 0.8035714285714286 in 2.1104354858398438s\n",
      "Finished epoch 72/100 with avg reward 0.7142857142857143 in 2.139005184173584s\n",
      "Finished epoch 73/100 with avg reward 0.8214285714285714 in 2.2268524169921875s\n",
      "Finished epoch 74/100 with avg reward 0.7857142857142857 in 2.0674829483032227s\n",
      "Finished epoch 75/100 with avg reward 0.8214285714285714 in 1.8486320972442627s\n",
      "Finished epoch 76/100 with avg reward 0.875 in 1.6807374954223633s\n",
      "Finished epoch 77/100 with avg reward 0.8214285714285714 in 1.7389135360717773s\n",
      "Finished epoch 78/100 with avg reward 0.8571428571428571 in 1.7677927017211914s\n",
      "Finished epoch 79/100 with avg reward 0.8928571428571429 in 1.7157905101776123s\n",
      "Finished epoch 80/100 with avg reward 0.8928571428571429 in 1.706132411956787s\n",
      "Finished epoch 81/100 with avg reward 0.8928571428571429 in 1.7236168384552002s\n",
      "Finished epoch 82/100 with avg reward 0.7857142857142857 in 1.7927894592285156s\n",
      "Finished epoch 83/100 with avg reward 0.8392857142857143 in 1.7154264450073242s\n",
      "Finished epoch 84/100 with avg reward 0.7857142857142857 in 1.724243402481079s\n",
      "Finished epoch 85/100 with avg reward 0.7678571428571429 in 1.6818046569824219s\n",
      "Finished epoch 86/100 with avg reward 0.75 in 1.71575927734375s\n",
      "Finished epoch 87/100 with avg reward 0.8035714285714286 in 1.7192895412445068s\n",
      "Finished epoch 88/100 with avg reward 0.7857142857142857 in 1.700697898864746s\n",
      "Finished epoch 89/100 with avg reward 0.8392857142857143 in 1.7331271171569824s\n",
      "Finished epoch 90/100 with avg reward 0.8571428571428571 in 1.6915478706359863s\n",
      "Finished epoch 91/100 with avg reward 0.8214285714285714 in 1.7116563320159912s\n",
      "Finished epoch 92/100 with avg reward 0.7857142857142857 in 1.705378770828247s\n",
      "Finished epoch 93/100 with avg reward 0.75 in 1.7074229717254639s\n",
      "Finished epoch 94/100 with avg reward 0.75 in 1.6866984367370605s\n",
      "Finished epoch 95/100 with avg reward 0.8214285714285714 in 1.7509431838989258s\n",
      "Finished epoch 96/100 with avg reward 0.8571428571428571 in 1.7261426448822021s\n",
      "Finished epoch 97/100 with avg reward 0.8035714285714286 in 1.763920545578003s\n",
      "Finished epoch 98/100 with avg reward 0.8392857142857143 in 1.7495906352996826s\n",
      "Finished epoch 99/100 with avg reward 0.8928571428571429 in 1.779637336730957s\n"
     ]
    }
   ],
   "source": [
    "avg_reward_stochastic_100 = ucb_stochastic_100.run(num_epochs=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x2603a067ef0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA0QAAAFNCAYAAAAgia/PAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2603a01eda0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rcParams['figure.figsize'] = 14, 5\n",
    "plt.plot(range(len(avg_reward_stochastic_100)), avg_reward_stochastic_100)\n",
    "plt.ylabel('average reward in epoch')\n",
    "plt.ylim(0.2, 1)\n",
    "plt.xlabel('epoch (1 epoch = recommended to all users)')\n",
    "plt.title('HybridLinUCB: Stochastic rewards with p=0.9')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}