Naive_Bayes_28.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from time import time\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   movie_id        movie title release date\n",
      "0         1   Toy Story (1995)  01-Jan-1995\n",
      "1         2   GoldenEye (1995)  01-Jan-1995\n",
      "2         3  Four Rooms (1995)  01-Jan-1995\n",
      "3         4  Get Shorty (1995)  01-Jan-1995\n",
      "4         5     Copycat (1995)  01-Jan-1995\n",
      "Here\n",
      "      movie_id                                        movie title release date\n",
      "674        675  Nosferatu (Nosferatu, eine Symphonie des Graue...   1922-01-01\n",
      "1541      1542                         Scarlet Letter, The (1926)   1926-01-01\n",
      "616        617          Blue Angel, The (Blaue Engel, Der) (1930)   1930-01-01\n",
      "655        656                                           M (1931)   1931-01-01\n",
      "1123      1124                         Farewell to Arms, A (1932)   1932-01-01\n",
      "Latest Year:  1998-10-23T00:00:00.000000000\n"
     ]
    }
   ],
   "source": [
    "#1622\n",
    "#User 564\n",
    "#Reading users file:\n",
    "u_cols = ['user_id','age','sex','occupation','zip_code']\n",
    "users = pd.read_csv('ml-100k/u.user', sep='|', \\\n",
    "                    names=u_cols,encoding='latin-1')\n",
    "\n",
    "#Reading ratings file:\n",
    "r_cols = ['user_id','movie_id','rating','unix_timestamp']\n",
    "rating = pd.read_csv('ml-100k/u.data',sep='\\t',names=r_cols,\\\n",
    "                     encoding='latin-1')\n",
    "\n",
    "#Reading the movie file:\n",
    "item_cols = ['movie_id','movie title','release date',\n",
    "'video release date','IMDb URL','unknown','Action',\n",
    "'Adventure','Animation','Childrens','Comedy','Crime',\n",
    "'Documentary','Drama','Fantasy','Film-Noir','Horror',\n",
    "'Musical','Mystery','Romance ','Sci-Fi','Thriller',\n",
    "'War' ,'Western']\n",
    "\n",
    "movies = pd.read_csv('ml-100k/u.item', sep='|',names=item_cols,\\\n",
    "                   encoding='latin-1')\n",
    "movies_prof = movies[['movie_id','movie title','release date']].copy()\n",
    "movies_prof1 = movies_prof.copy()\n",
    "print(movies_prof1.head())\n",
    "movies_prof1['release date'] = pd.to_datetime(movies_prof1['release date'])\n",
    "movies_prof1 = movies_prof1.sort_values(['release date'])\n",
    "#print(movies_prof1)\n",
    "profit = pd.DataFrame(columns=['movie_id','Profit'])\n",
    "print(\"Here\")\n",
    "print(movies_prof1.head())\n",
    "#early = int(str(movies_prof1['release date'].iloc[-1])[7:11])\n",
    "last = \"NaT\"\n",
    "j = -1\n",
    "while(last == \"NaT\"):\n",
    "    last = str(movies_prof1['release date'].values[j])\n",
    "    j-=1\n",
    "print(\"Latest Year: \", last)\n",
    "last = int(last[:4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   user_id  movie_id  rating  unix_timestamp\n",
      "0      196       242       3       881250949\n",
      "1      186       302       3       891717742\n",
      "2       22       377       1       878887116\n",
      "3      244        51       2       880606923\n",
      "4      166       346       1       886397596\n",
      "right here\n",
      "      user_id age  sex movie_id\n",
      "0           1  24    1       61\n",
      "1           1  24    1      189\n",
      "2           1  24    1       33\n",
      "3           1  24    1      160\n",
      "4           1  24    1       20\n",
      "5           1  24    1      202\n",
      "6           1  24    1      171\n",
      "7           1  24    1      265\n",
      "8           1  24    1      117\n",
      "9           1  24    1       47\n",
      "10          1  24    1      222\n",
      "11          1  24    1      253\n",
      "12          1  24    1      113\n",
      "13          1  24    1      227\n",
      "14          1  24    1       17\n",
      "15          1  24    1       90\n",
      "16          1  24    1       64\n",
      "17          1  24    1       92\n",
      "18          1  24    1      228\n",
      "19          1  24    1      121\n",
      "20          1  24    1      114\n",
      "21          1  24    1      132\n",
      "22          1  24    1      134\n",
      "23          1  24    1       98\n",
      "24          1  24    1      186\n",
      "25          1  24    1      221\n",
      "26          1  24    1       84\n",
      "27          1  24    1       31\n",
      "28          1  24    1       70\n",
      "29          1  24    1       60\n",
      "...       ...  ..  ...      ...\n",
      "52364     562  54    0      566\n",
      "52365     562  54    0      432\n",
      "52366     562  54    0      234\n",
      "52367     562  54    0      485\n",
      "52368     562  54    0       82\n",
      "52369     563  39    0     1035\n",
      "52370     563  39    0      692\n",
      "52371     563  39    0      233\n",
      "52372     563  39    0      401\n",
      "52373     563  39    0      367\n",
      "52374     563  39    0      254\n",
      "52375     563  39    0       50\n",
      "52376     563  39    0      237\n",
      "52377     563  39    0      210\n",
      "52378     563  39    0      118\n",
      "52379     563  39    0      153\n",
      "52380     563  39    0      181\n",
      "52381     563  39    0      781\n",
      "52382     563  39    0      403\n",
      "52383     563  39    0       70\n",
      "52384     563  39    0      476\n",
      "52385     563  39    0      255\n",
      "52386     563  39    0      257\n",
      "52387     563  39    0      321\n",
      "52388     563  39    0      301\n",
      "52389     563  39    0      172\n",
      "52390     563  39    0      167\n",
      "52391     563  39    0      220\n",
      "52392     563  39    0      294\n",
      "52393     563  39    0      566\n",
      "\n",
      "[52394 rows x 4 columns]\n"
     ]
    }
   ],
   "source": [
    "print(rating.head())\n",
    "df2  = rating.copy()\n",
    "\n",
    "df2.head()\n",
    "\n",
    "groups = df2.groupby(\"user_id\", as_index=False)\n",
    "k = 0\n",
    "columns = ['user_id','age','sex','movie_id']\n",
    "#print(users)\n",
    "df_users = pd.DataFrame(columns=columns)\n",
    "#print(df_user)\n",
    "for group in groups:\n",
    "    toy_stor = group\n",
    "    toy_story = toy_stor[1].copy()\n",
    "    #print((toy_story) )\n",
    "    movie_num = toy_story['movie_id'].iloc[0]\n",
    "    users_rating = users.copy()\n",
    "    df_1_toy = users.copy()\n",
    "    #df_1_toy.occupation.unique()\n",
    "\n",
    "    user_movie = pd.merge(df_1_toy,toy_story)\n",
    "    #print(\"Sigh\\n\",user_movie)\n",
    "    \n",
    "\n",
    "    #user_movie = user_movie.drop(['unix_timestamp','rating','zip_code'],axis=1)\n",
    "    #print(user_movie.head())\n",
    "    user_movie = user_movie.loc[user_movie['rating'] >= 3]\n",
    "    #print(user_movie.head())\n",
    "    user_movie = user_movie.drop(['unix_timestamp','rating','zip_code','occupation'],axis=1)\n",
    "    \n",
    "    df_users = df_users.append(user_movie)\n",
    "        \n",
    "    k+=1\n",
    "\n",
    "print(\"right here\")\n",
    "df_users = df_users.reset_index(drop=True)\n",
    "size = df_users.loc[df_users['user_id'] < 564].shape[0]\n",
    "df_users = df_users[:size].copy()\n",
    "df_users.sex.replace(['M', 'F'], [1, 0], inplace=True)\n",
    "print(df_users)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "labels_train\n",
      "\n",
      "        movie_id\n",
      "user_id         \n",
      "1             61\n",
      "1            189\n",
      "1             33\n",
      "1            160\n",
      "1             20\n",
      "sort:\n",
      "user_id\n",
      "340    1\n",
      "65     1\n",
      "493    1\n",
      "445    1\n",
      "291    1\n",
      "Name: movie_id, dtype: int64\n",
      "      user_id age  sex\n",
      "30753     340  46    1\n",
      "5920       65  51    0\n",
      "45714     493  22    1\n",
      "40909     445  21    1\n",
      "24275     291  19    1\n",
      "38939     422  26    1\n",
      "40159     435  24    1\n",
      "11329     130  20    1\n",
      "19777     247  28    1\n",
      "20527     256  35    0\n",
      "8238       92  32    1\n",
      "1687       13  47    1\n",
      "23234     280  30    0\n",
      "35431     390  42    0\n",
      "30598     339  35    1\n",
      "35387     389  44    0\n",
      "16387     203  25    0\n",
      "4466       54  22    1\n",
      "6848       82  50    1\n",
      "16372     202  41    0\n",
      "43410     467  29    1\n",
      "13171     160  27    1\n",
      "51172     550  16    0\n",
      "39957     434  16    0\n",
      "8967       97  43    1\n",
      "11386     131  59    0\n",
      "36922     402  30    1\n",
      "47701     514  27    1\n",
      "16711     209  33    0\n",
      "10550     121  54    1\n",
      "...       ...  ..  ...\n",
      "44987     486  39    1\n",
      "48022     519  22    1\n",
      "40682     440  30    1\n",
      "40878     445  21    1\n",
      "48015     519  22    1\n",
      "37775     409  48    1\n",
      "49170     532  20    1\n",
      "38503     416  20    0\n",
      "39142     425  19    1\n",
      "45055     486  39    1\n",
      "43197     464  60    1\n",
      "40517     437  27    0\n",
      "47828     514  27    1\n",
      "40664     439  23    0\n",
      "41064     448  23    1\n",
      "41313     450  35    0\n",
      "42479     456  24    1\n",
      "43364     466  22    1\n",
      "45057     486  39    1\n",
      "45083     486  39    1\n",
      "48033     519  22    1\n",
      "45475     489  55    1\n",
      "45492     489  55    1\n",
      "46022     496  21    0\n",
      "46267     497  20    1\n",
      "46527     500  28    1\n",
      "48007     519  22    1\n",
      "50619     543  33    1\n",
      "51477     552  45    1\n",
      "51167     550  16    0\n",
      "\n",
      "[52394 rows x 3 columns]\n",
      "here\n",
      "Labels size 52394\n",
      "      user_id age  sex\n",
      "30753     340  46    1\n",
      "5920       65  51    0\n",
      "45714     493  22    1\n",
      "40909     445  21    1\n",
      "24275     291  19    1\n",
      "        age  sex\n",
      "user_id         \n",
      "340      46    1\n",
      "65       51    0\n",
      "493      22    1\n",
      "445      21    1\n",
      "291      19    1\n"
     ]
    }
   ],
   "source": [
    "labels_train = df_users[['movie_id','user_id']].copy()\n",
    "labels_train = labels_train.set_index('user_id')\n",
    "#print(labels_train.head())\n",
    "print(\"\\nlabels_train\\n\")\n",
    "print(labels_train.head())\n",
    "print(\"sort:\")\n",
    "labels_train = pd.Series(df_users['movie_id'].values,\\\n",
    "                         index=df_users['user_id'])\n",
    "labels_train = labels_train.rename(\"movie_id\")\n",
    "labels_train= labels_train.\\\n",
    "apply(pd.to_numeric, errors='coerce')\n",
    "labels_train = labels_train.sort_values()\n",
    "print(labels_train.head())\n",
    "df_user = df_users.copy()\n",
    "df_user = df_user.sort_values(by=['movie_id'])\n",
    "df_user = df_user.drop(['movie_id'],axis=1)\n",
    "#print(df_users.set_index('user_id'))\n",
    "#df_users = df_users.set_index('user_id')\n",
    "print(df_user)\n",
    "print(\"here\")\n",
    "print(\"Labels size\",len(labels_train))\n",
    "#print(labels_train.loc[labels_train['movie_id'] == 586].shape[0])\n",
    "print(df_user.head())\n",
    "df_user = df_user.set_index('user_id')\n",
    "print(df_user.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "movies1 = labels_train.unique()\n",
    "movies1 = movies1.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user_id\n",
      "340    1\n",
      "65     1\n",
      "493    1\n",
      "445    1\n",
      "291    1\n",
      "Name: movie_id, dtype: int64\n",
      "user_id\n",
      "500    1616\n",
      "519    1617\n",
      "543    1619\n",
      "552    1620\n",
      "550    1620\n",
      "Name: movie_id, dtype: int64\n",
      "        age  sex\n",
      "user_id         \n",
      "340      46    1\n",
      "65       51    0\n",
      "493      22    1\n",
      "445      21    1\n",
      "291      19    1\n",
      "        age  sex\n",
      "user_id         \n",
      "500      28    1\n",
      "519      22    1\n",
      "543      33    1\n",
      "552      45    1\n",
      "550      16    0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "GNB = MultinomialNB()\n",
    "t0 = time()\n",
    "print(labels_train.head())\n",
    "print(labels_train.tail())\n",
    "print(df_user.head())\n",
    "print(df_user.tail())\n",
    "GNB.fit(df_user,labels_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predictions\n",
    "This goes about predicting the movies that users should watch using the Multinomial Naive Bayes Model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#for i in [563,566,567,568]:\n",
    "#print(movies1[:10])\n",
    "genre = ['release date','video release date','IMDb URL','unknown','Action',\\\n",
    "            'Adventure','Animation','Childrens','Comedy','Crime',\\\n",
    "            'Documentary','Drama','Fantasy','Film-Noir','Horror',\\\n",
    "            'Musical','Mystery','Romance ','Sci-Fi','Thriller',\\\n",
    "            'War' ,'Western']\n",
    "#valid_users = [564, 566, 567, 568, 569, 570, 571, 572, 573, 575, 576, 577, 578, 579, 580, 581, 582, 583, 584, 586, 588, 589, 590, 591, 593, 594, 595, 596, 597, 598, 599, 600, 601, 602, 603, 604, 605, 606, 607, 608, 610, 611, 612, 613, 614, 615, 616, 617, 618, 619, 620, 621, 623, 624, 625, 626, 627, 628, 629, 630, 631, 632, 633, 634, 635, 636, 638, 639, 640, 641, 643, 645, 646, 647, 649, 651, 652, 653, 654, 656, 657, 658, 659, 660, 661, 664, 665, 666, 667, 668, 669, 670, 671, 672, 673, 674, 677, 678, 679, 680, 683, 684, 685, 686, 687, 688, 689, 690, 691, 692, 693, 695, 696, 697, 698, 700, 701, 702, 703, 704, 706, 709, 710, 711, 712, 714, 715, 716, 717, 718, 719, 720, 721, 722, 723, 725, 726, 728, 729, 730, 731, 732, 734, 735, 736, 737, 738, 740, 741, 742, 743, 744, 745, 746, 748, 749, 750, 752, 753, 754, 755, 757, 759, 760, 763, 764, 765, 766, 767, 768, 769, 770, 771, 772, 773, 774, 775, 777, 778, 779, 780, 781, 783, 784, 785, 786, 788, 789, 790, 791, 792, 794, 795, 796, 797, 800, 801, 802, 803, 806, 807, 808, 809, 810, 811, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 825, 826, 827, 829, 830, 831, 832, 834, 836, 837, 838, 841, 843, 844, 848, 849, 850, 852, 853, 855, 856, 857, 858, 859, 860, 861, 862, 865, 866, 872, 873, 874, 875, 876, 877, 878, 879, 881, 884, 885, 886, 888, 889, 890, 891, 892, 893, 895, 897, 898, 899, 900, 902, 903, 904, 905, 906, 907, 908, 909, 910, 911, 912, 913, 914, 915, 917, 919, 920, 922, 923, 924, 926, 927, 928, 929, 930, 931, 933, 934, 935, 937, 938, 939, 940, 941, 942, 943]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1483\n"
     ]
    }
   ],
   "source": [
    "groups = df2.groupby(\"user_id\", as_index=False)\n",
    "#print(movies1)\n",
    "k=1\n",
    "user_num=0\n",
    "index = user_num\n",
    "print(len(movies1))\n",
    "valid_users = []\n",
    "jk = 1\n",
    "for group in groups:\n",
    "    if(jk > 563):\n",
    "        toy_stor = group\n",
    "        toy_story = toy_stor[1].copy()\n",
    "        if(set(toy_story['movie_id'].tolist()).issubset(movies1) == False):\n",
    "            k+=1\n",
    "            continue\n",
    "        else:\n",
    "            valid_users.append(group[0])\n",
    "    jk+=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Movies sorted:  1483\n",
      "306\n",
      "For  1  users:\n",
      "        age sex\n",
      "user_id        \n",
      "605      33   1 \n",
      "\n",
      "Probability:\n",
      "0                    Star Wars (1977)\n",
      "1                        Fargo (1996)\n",
      "2           Return of the Jedi (1983)\n",
      "3                      Contact (1997)\n",
      "4                       Scream (1996)\n",
      "5                    Toy Story (1995)\n",
      "6      Raiders of the Lost Ark (1981)\n",
      "7    Silence of the Lambs, The (1991)\n",
      "8                 Pulp Fiction (1994)\n",
      "9               Twelve Monkeys (1995)\n",
      "Name: movie title, dtype: object\n",
      "Short term:\n",
      "0                            Star Wars (1977)\n",
      "53                   Wizard of Oz, The (1939)\n",
      "63                          Casablanca (1942)\n",
      "10                      Godfather, The (1972)\n",
      "6              Raiders of the Lost Ark (1981)\n",
      "78               It's a Wonderful Life (1946)\n",
      "102                       Citizen Kane (1941)\n",
      "23     Monty Python and the Holy Grail (1974)\n",
      "13            Empire Strikes Back, The (1980)\n",
      "64                              Psycho (1960)\n",
      "Name: movie title, dtype: object\n",
      "Long term\n",
      "0                            Star Wars (1977)\n",
      "53                   Wizard of Oz, The (1939)\n",
      "63                          Casablanca (1942)\n",
      "10                      Godfather, The (1972)\n",
      "6              Raiders of the Lost Ark (1981)\n",
      "78               It's a Wonderful Life (1946)\n",
      "102                       Citizen Kane (1941)\n",
      "23     Monty Python and the Holy Grail (1974)\n",
      "13            Empire Strikes Back, The (1980)\n",
      "64                              Psycho (1960)\n",
      "Name: movie title, dtype: object\n",
      "For most probable movie\n",
      "Average for most probabile movie:  0.051304603693488764\n",
      "Average for movie that maximizes P(i) * R(i) : 0.03640809443192415\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.03640809443192415\n",
      "For maximum short-term profit\n",
      "Average for most probabile movie:  0.43962107983912035\n",
      "Average for movie that maximizes P(i) * R(i) : 0.5493739316378934\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.5493739316378934\n",
      "True\n",
      "False\n",
      "For maximum long-term profit\n",
      "Average for most probabile movie:  0.44194991777891857\n",
      "Average for movie that maximizes P(i) * R(i) : 0.5514654414965147\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.5514654414965147\n",
      "True\n",
      "False\n",
      "\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#valid_users = [564,566,567,568,569]\n",
    "print(\"Movies sorted: \",len(movies1))\n",
    "#for i in [563,566,567]:\n",
    "#for i in range(563,943):\n",
    "#for i in valid_users:\n",
    "prob = [0,0,0]\n",
    "st_prof = [0,0,0]\n",
    "lt_prof = [0,0,0]\n",
    "#print(temp[:5])\n",
    "temp = valid_users.copy()\n",
    "random.shuffle(temp)\n",
    "metrics = ['Probability','P(i)*R(i)','(P(i)*R(i)/(1-P(i))']\n",
    "print(len(temp))\n",
    "length_users = 1\n",
    "print(\"For \",length_users,\" users:\")\n",
    "for i in temp[:length_users]:\n",
    "    prob = [0,0,0]\n",
    "    st_prof = [0,0,0]\n",
    "    lt_prof = [0,0,0]\n",
    "    i-=1\n",
    "    #print(\"yeah\")\n",
    "    df_sugg = pd.DataFrame(columns = ['user_id','age','sex'])\n",
    "    df_user = users.iloc[i].copy()\n",
    "    df_user = df_user.drop(['occupation','zip_code'], axis=0)\n",
    "    #print(df_user)\n",
    "    #print(type(df_user))\n",
    "    #df_user = df_user.set_index('user_id')\n",
    "    if(df_user['sex'] == 'M'):\n",
    "        df_user['sex'] = 1\n",
    "    else:\n",
    "        df_user['sex'] = 0\n",
    "    df_sugg = df_sugg.append(df_user)\n",
    "    #print(df_sugg)\n",
    "    df_sugg = df_sugg.set_index('user_id')\n",
    "    nb_pred = GNB.predict(df_sugg)\n",
    "    approx_pred = GNB.predict_proba(df_sugg)\n",
    "    proba = approx_pred[0].copy()\n",
    "    proba1 = proba.copy()\n",
    "    mov = pd.DataFrame(proba,columns=['Probability'])\n",
    "    mov['movie_id'] = None\n",
    "    for j in range(len(movies1)):\n",
    "        mov.loc[j,'movie_id'] = movies1[j]\n",
    "    mov = mov.sort_values(['Probability'],ascending=False)\n",
    "    combine = pd.merge(mov,movies_prof1,how='outer')\n",
    "    combine = combine.dropna()\n",
    "    combine = combine.reset_index(drop=True)\n",
    "    k = 2\n",
    "    #print(combine['release date'].astype(str)[:][:4])\n",
    "    combine['year'] = combine['release date'].astype(str).str.slice(0,4)\n",
    "    combine['year'] = combine['year'].astype(int)\n",
    "    combine['R'] = k + ( ((datetime.datetime.now()).year - combine['year']) / 4 )\n",
    "    combine['P(i)*R(i)'] = combine['Probability'] * combine['R']\n",
    "    combine['(P(i)*R(i)/(1-P(i))'] = combine['P(i)*R(i)'] / (1 - combine['Probability'])\n",
    "    y=0\n",
    "    #i == 600 i == 700 i ==800 or \n",
    "    #if( i == 600 or i == 700 or i ==800 or i == 901):\n",
    "    #if(i%10 == 0):\n",
    "    #    print(df_sugg, \"\\n\")\n",
    "    print(df_sugg, \"\\n\")\n",
    "    print(\"Probability:\")\n",
    "    print((combine.sort_values('Probability',ascending=False)[:10]['movie title']))\n",
    "    print(\"Short term:\")\n",
    "    print((combine.sort_values('P(i)*R(i)',ascending=False)[:10]['movie title']))\n",
    "    print(\"Long term\")\n",
    "    print((combine.sort_values('(P(i)*R(i)/(1-P(i))',ascending=False)[:10]['movie title']))\n",
    "    for jk in metrics:\n",
    "        P_acc = (combine.sort_values('Probability',ascending=False))[jk]\n",
    "        P_st = (combine.sort_values('P(i)*R(i)',ascending=False))[jk]\n",
    "        P_lt = (combine.sort_values('(P(i)*R(i)/(1-P(i))',ascending=False))[jk]\n",
    "        if(jk == 'Probability'):            \n",
    "            prob[0]+=P_acc[:10].sum()\n",
    "            prob[1]+=P_st[:10].sum()\n",
    "            prob[2]+=P_lt[:10].sum()\n",
    "        elif(jk == 'P(i)*R(i)'):\n",
    "            st_prof[0]+=P_acc[:10].sum()\n",
    "            st_prof[1]+=P_st[:10].sum()\n",
    "            st_prof[2]+=P_lt[:10].sum()\n",
    "        else:\n",
    "            lt_prof[0]+=P_acc[:10].sum()\n",
    "            lt_prof[1]+=P_st[:10].sum()\n",
    "            lt_prof[2]+=P_lt[:10].sum()    \n",
    "\n",
    "    len_valid_users  = length_users\n",
    "    avg_prob_max_Pi = prob[0] / len_valid_users\n",
    "    print(\"For most probable movie\")\n",
    "    print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "    avg_prob_st_Rev = prob[1] / len_valid_users\n",
    "    print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "    avg_prob_lt_Rev = prob[2] / len_valid_users\n",
    "    print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "          avg_prob_lt_Rev)\n",
    "\n",
    "    print(\"For maximum short-term profit\")\n",
    "    len_valid_users  = length_users\n",
    "    avg_prob_max_Pi = st_prof[0] / len_valid_users\n",
    "    print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "    avg_prob_st_Rev = st_prof[1] / len_valid_users\n",
    "    print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "    avg_prob_lt_Rev = st_prof[2] / len_valid_users\n",
    "    print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "          avg_prob_lt_Rev)\n",
    "\n",
    "    print(avg_prob_st_Rev == avg_prob_lt_Rev)\n",
    "    print(avg_prob_st_Rev > avg_prob_lt_Rev)\n",
    "\n",
    "\n",
    "    print(\"For maximum long-term profit\")\n",
    "    len_valid_users  = length_users\n",
    "    avg_prob_max_Pi = lt_prof[0] / len_valid_users\n",
    "    print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "    avg_prob_st_Rev = lt_prof[1] / len_valid_users\n",
    "    print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "    avg_prob_lt_Rev = lt_prof[2] / len_valid_users\n",
    "    print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "          avg_prob_lt_Rev)\n",
    "\n",
    "    print(avg_prob_st_Rev == avg_prob_lt_Rev)\n",
    "    print(avg_prob_st_Rev < avg_prob_lt_Rev)\n",
    "    print(\"\\n\\n\\n\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Movies sorted:  1483\n",
      "306\n",
      "For  306  users:\n",
      "        age sex\n",
      "user_id        \n",
      "581      37   1 \n",
      "\n",
      "        age sex\n",
      "user_id        \n",
      "777      63   1 \n",
      "\n",
      "        age sex\n",
      "user_id        \n",
      "646      17   0 \n",
      "\n",
      "        age sex\n",
      "user_id        \n",
      "688      37   0 \n",
      "\n",
      "        age sex\n",
      "user_id        \n",
      "742      35   1 \n",
      "\n",
      "        age sex\n",
      "user_id        \n",
      "638      45   1 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "#valid_users = [564,566,567,568,569]\n",
    "print(\"Movies sorted: \",len(movies1))\n",
    "#for i in [563,566,567]:\n",
    "#for i in range(563,943):\n",
    "#for i in valid_users:\n",
    "prob = [0,0,0]\n",
    "st_prof = [0,0,0]\n",
    "lt_prof = [0,0,0]\n",
    "temp = valid_users.copy()\n",
    "#print(temp[:5])\n",
    "random.shuffle(temp)\n",
    "metrics = ['Probability','P(i)*R(i)','(P(i)*R(i)/(1-P(i))']\n",
    "print(len(temp))\n",
    "print(\"For \",len(valid_users),\" users:\")\n",
    "xy = 0\n",
    "for i in temp[:len(valid_users)]:\n",
    "    xy+=1\n",
    "    i-=1\n",
    "    #print(\"yeah\")\n",
    "    df_sugg = pd.DataFrame(columns = ['user_id','age','sex'])\n",
    "    df_user = users.iloc[i].copy()\n",
    "    df_user = df_user.drop(['occupation','zip_code'], axis=0)\n",
    "    #print(df_user)\n",
    "    #print(type(df_user))\n",
    "    #df_user = df_user.set_index('user_id')\n",
    "    if(df_user['sex'] == 'M'):\n",
    "        df_user['sex'] = 1\n",
    "    else:\n",
    "        df_user['sex'] = 0\n",
    "    df_sugg = df_sugg.append(df_user)\n",
    "    #print(df_sugg)\n",
    "    df_sugg = df_sugg.set_index('user_id')\n",
    "    nb_pred = GNB.predict(df_sugg)\n",
    "    approx_pred = GNB.predict_proba(df_sugg)\n",
    "    proba = approx_pred[0].copy()\n",
    "    proba1 = proba.copy()\n",
    "    mov = pd.DataFrame(proba,columns=['Probability'])\n",
    "    mov['movie_id'] = None\n",
    "    for j in range(len(movies1)):\n",
    "        mov.loc[j,'movie_id'] = movies1[j]\n",
    "    mov = mov.sort_values(['Probability'],ascending=False)\n",
    "    combine = pd.merge(mov,movies_prof1,how='outer')\n",
    "    combine = combine.dropna()\n",
    "    combine = combine.reset_index(drop=True)\n",
    "    k = 2\n",
    "    #print(combine['release date'].astype(str)[:][:4])\n",
    "    combine['year'] = combine['release date'].astype(str).str.slice(0,4)\n",
    "    combine['year'] = combine['year'].astype(int)\n",
    "    combine['R'] = k + ( ((datetime.datetime.now()).year - combine['year']) / 4 )\n",
    "    combine['P(i)*R(i)'] = combine['Probability'] * combine['R']\n",
    "    combine['(P(i)*R(i)/(1-P(i))'] = combine['P(i)*R(i)'] / (1 - combine['Probability'])\n",
    "    y=0\n",
    "    #i == 600 i == 700 i ==800 or \n",
    "    #if( i == 600 or i == 700 or i ==800 or i == 901):\n",
    "    if(xy%50 == 0):\n",
    "        print(df_sugg, \"\\n\")\n",
    "    for jk in metrics:\n",
    "        P_acc = (combine.sort_values('Probability',ascending=False))[jk]\n",
    "        P_st = (combine.sort_values('P(i)*R(i)',ascending=False))[jk]\n",
    "        P_lt = (combine.sort_values('(P(i)*R(i)/(1-P(i))',ascending=False))[jk]\n",
    "        if(jk == 'Probability'):            \n",
    "            prob[0]+=P_acc[:10].sum()\n",
    "            prob[1]+=P_st[:10].sum()\n",
    "            prob[2]+=P_lt[:10].sum()\n",
    "        elif(jk == 'P(i)*R(i)'):\n",
    "            st_prof[0]+=P_acc[:10].sum()\n",
    "            st_prof[1]+=P_st[:10].sum()\n",
    "            st_prof[2]+=P_lt[:10].sum()\n",
    "        else:\n",
    "            lt_prof[0]+=P_acc[:10].sum()\n",
    "            lt_prof[1]+=P_st[:10].sum()\n",
    "            lt_prof[2]+=P_lt[:10].sum() \n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For most probable movie\n",
      "Average for most probabile movie:  0.05169425308334061\n",
      "Average for movie that maximizes P(i) * R(i) : 0.03578611023604978\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.035798180172099694\n",
      "For maximum short-term profit\n",
      "Average for most probabile movie:  0.4535311742032724\n",
      "Average for movie that maximizes P(i) * R(i) : 0.5578120260846706\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.5578115592145294\n",
      "False\n",
      "True\n",
      "For maximum long-term profit\n",
      "Average for most probabile movie:  0.45594131582083564\n",
      "Average for movie that maximizes P(i) * R(i) : 0.5599046660017819\n",
      "Average for movie that maximizes (P(i)*R(i)/(1-P(i)) :  0.5599047463328457\n",
      "False\n",
      "True\n",
      "\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "len_valid_users  = len(valid_users)\n",
    "avg_prob_max_Pi = prob[0] / len_valid_users\n",
    "print(\"For most probable movie\")\n",
    "print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "avg_prob_st_Rev = prob[1] / len_valid_users\n",
    "print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "avg_prob_lt_Rev = prob[2] / len_valid_users\n",
    "print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "      avg_prob_lt_Rev)\n",
    "\n",
    "print(\"For maximum short-term profit\")\n",
    "len_valid_users  = len(valid_users)\n",
    "avg_prob_max_Pi = st_prof[0] / len_valid_users\n",
    "print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "avg_prob_st_Rev = st_prof[1] / len_valid_users\n",
    "print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "avg_prob_lt_Rev = st_prof[2] / len_valid_users\n",
    "print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "      avg_prob_lt_Rev)\n",
    "\n",
    "print(avg_prob_st_Rev == avg_prob_lt_Rev)\n",
    "print(avg_prob_st_Rev > avg_prob_lt_Rev)\n",
    "\n",
    "\n",
    "print(\"For maximum long-term profit\")\n",
    "len_valid_users  = len(valid_users)\n",
    "avg_prob_max_Pi = lt_prof[0] / len_valid_users\n",
    "print(\"Average for most probabile movie: \",avg_prob_max_Pi)\n",
    "avg_prob_st_Rev = lt_prof[1] / len_valid_users\n",
    "print(\"Average for movie that maximizes P(i) * R(i) :\",avg_prob_st_Rev)\n",
    "avg_prob_lt_Rev = lt_prof[2] / len_valid_users\n",
    "print(\"Average for movie that maximizes (P(i)*R(i)/(1-P(i)) : \",\\\n",
    "      avg_prob_lt_Rev)\n",
    "\n",
    "print(avg_prob_st_Rev == avg_prob_lt_Rev)\n",
    "print(avg_prob_st_Rev < avg_prob_lt_Rev)\n",
    "print(\"\\n\\n\\n\")"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}