10_JokeRatingPrediction/JokesRatingPrediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "## Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:08.993912Z",
     "start_time": "2021-01-15T00:34:07.588592Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from fastai.tabular.all import *\n",
    "from fastai.collab import *"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "## Read & Explore Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:12.085591Z",
     "start_time": "2021-01-15T00:34:11.268246Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv(\"../data/JokeRatingPredictions/train.csv\")\n",
    "test = pd.read_csv(\"../data/JokeRatingPredictions/test.csv\")\n",
    "jokes = pd.read_csv(\"../data/JokeRatingPredictions/jokes.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:12.098742Z",
     "start_time": "2021-01-15T00:34:12.087978Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1092059, 4),\n",
       "           id  user_id  joke_id  Rating\n",
       " 0  31030_110    31030      110   2.750\n",
       " 1  16144_109    16144      109   5.094\n",
       " 2    23098_6    23098        6  -6.438\n",
       " 3   14273_86    14273       86   4.406\n",
       " 4  18419_134    18419      134   9.375,\n",
       " (537880, 3),\n",
       "           id  user_id  joke_id\n",
       " 0    6194_11     6194       11\n",
       " 1    19356_3    19356        3\n",
       " 2   23426_79    23426       79\n",
       " 3    40030_3    40030        3\n",
       " 4  19806_115    19806      115,\n",
       " (139, 2),\n",
       "    joke_id  \\\n",
       " 0        1   \n",
       " 1        2   \n",
       " 2        3   \n",
       " 3        4   \n",
       " 4        5   \n",
       " \n",
       "                                                                                              joke_text  \n",
       " 0             Q. What's O. J. Simpson's web address? A. Slash, slash, backslash, slash, slash, escape.  \n",
       " 1                          How many feminists does it take to screw in a light bulb? That's not funny.  \n",
       " 2                    Q. Did you hear about the dyslexic devil worshiper? A. He sold his soul to Santa.  \n",
       " 3  They asked the Japanese visitor if they have elections in his country. \"Every morning,\" he answers.  \n",
       " 4                 Q: What did the blind person say when given some matzah? A: Who the hell wrote this?  )"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape, train.head(), test.shape, test.head(), jokes.shape, jokes.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:13.244084Z",
     "start_time": "2021-01-15T00:34:13.227638Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(40863, 40863)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# ~41K users, 1MM interactions\n",
    "train.user_id.nunique(), test.user_id.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:14.199686Z",
     "start_time": "2021-01-15T00:34:13.907789Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='Frequency'>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train.Rating.plot(kind='hist')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "hidden": true
   },
   "source": [
    "Ratings range from -10 to 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:34:41.558633Z",
     "start_time": "2021-01-15T00:34:35.709204Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>joke_id</th>\n",
       "      <th>Rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4285</td>\n",
       "      <td>43</td>\n",
       "      <td>-7.906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>39998</td>\n",
       "      <td>9</td>\n",
       "      <td>6.719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>31226</td>\n",
       "      <td>99</td>\n",
       "      <td>5.781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40838</td>\n",
       "      <td>43</td>\n",
       "      <td>10.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5935</td>\n",
       "      <td>135</td>\n",
       "      <td>-5.688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>38849</td>\n",
       "      <td>101</td>\n",
       "      <td>6.219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>24775</td>\n",
       "      <td>98</td>\n",
       "      <td>0.531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>26597</td>\n",
       "      <td>108</td>\n",
       "      <td>2.031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>25143</td>\n",
       "      <td>31</td>\n",
       "      <td>-2.031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>14818</td>\n",
       "      <td>96</td>\n",
       "      <td>-4.375</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dls = CollabDataLoaders.from_df(\n",
    "    train,\n",
    "    user_name='user_id', \n",
    "    item_name='joke_id', \n",
    "    rating_name= \"Rating\",\n",
    "    bs=2**12,\n",
    "    valid_pct=0.2\n",
    ")\n",
    "dls.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:35:15.528411Z",
     "start_time": "2021-01-15T00:35:15.524223Z"
    }
   },
   "outputs": [],
   "source": [
    "class RMSELoss(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.mse = nn.MSELoss()\n",
    "        \n",
    "    def forward(self,yhat,y):\n",
    "        return torch.sqrt(self.mse(yhat,y))\n",
    "    \n",
    "def RMSELossFlat(*args, axis=-1, floatify=True, **kwargs):\n",
    "    \"Same as `nn.MSELoss`, but flattens input and target.\"\n",
    "    return BaseLoss(RMSELoss, *args, axis=axis, floatify=floatify, is_2d=False, **kwargs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:35:22.154912Z",
     "start_time": "2021-01-15T00:35:22.027621Z"
    }
   },
   "outputs": [],
   "source": [
    "learn = collab_learner(\n",
    "    dls, \n",
    "    n_factors=200, \n",
    "    use_nn=False,\n",
    "    y_range=(-10.5 ,10.5),\n",
    "    loss_func=RMSELossFlat(),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:36:31.048501Z",
     "start_time": "2021-01-15T00:35:34.512428Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>5.465502</td>\n",
       "      <td>5.440821</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>4.848229</td>\n",
       "      <td>4.506827</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.142750</td>\n",
       "      <td>4.144625</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>3.849428</td>\n",
       "      <td>4.082230</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>3.567850</td>\n",
       "      <td>4.056830</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>3.263375</td>\n",
       "      <td>4.042928</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>3.012081</td>\n",
       "      <td>4.037589</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>2.828375</td>\n",
       "      <td>4.036174</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>2.721868</td>\n",
       "      <td>4.036244</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>2.678576</td>\n",
       "      <td>4.036256</td>\n",
       "      <td>00:05</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.fit_one_cycle(10, 1e-3, wd=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So we get a good model with 4.036 RMSE with 20% Training Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training the final model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we need to train a model with all the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:37:43.990581Z",
     "start_time": "2021-01-15T00:37:39.816792Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>joke_id</th>\n",
       "      <th>Rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>22801</td>\n",
       "      <td>64</td>\n",
       "      <td>1.312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2031</td>\n",
       "      <td>43</td>\n",
       "      <td>6.188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>32972</td>\n",
       "      <td>104</td>\n",
       "      <td>3.219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>24263</td>\n",
       "      <td>6</td>\n",
       "      <td>-4.938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17531</td>\n",
       "      <td>8</td>\n",
       "      <td>-9.781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>15321</td>\n",
       "      <td>46</td>\n",
       "      <td>8.625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>629</td>\n",
       "      <td>34</td>\n",
       "      <td>-0.250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>12751</td>\n",
       "      <td>19</td>\n",
       "      <td>-2.938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>39016</td>\n",
       "      <td>7</td>\n",
       "      <td>9.406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>21766</td>\n",
       "      <td>2</td>\n",
       "      <td>8.281</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dls = CollabDataLoaders.from_df(\n",
    "    train,\n",
    "    user_name='user_id', \n",
    "    item_name='joke_id', \n",
    "    rating_name= \"Rating\",\n",
    "    bs=2**12,\n",
    "    valid_pct=0\n",
    ")\n",
    "dls.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:37:55.632305Z",
     "start_time": "2021-01-15T00:37:55.508231Z"
    }
   },
   "outputs": [],
   "source": [
    "learn = collab_learner(\n",
    "    dls, \n",
    "    n_factors=200, \n",
    "    use_nn=False,\n",
    "    y_range=(-10.5 ,10.5),\n",
    "    loss_func=RMSELossFlat(),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:39:09.817861Z",
     "start_time": "2021-01-15T00:38:03.788427Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>5.453224</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>4.549985</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.055283</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>3.829326</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>3.552691</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>3.229800</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>2.974170</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>2.785772</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>2.676441</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>2.644196</td>\n",
       "      <td>None</td>\n",
       "      <td>00:06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/aaagraw/anaconda3/envs/fastainew/lib/python3.8/site-packages/fastprogress/fastprogress.py:74: UserWarning: Your generator is empty.\n",
      "  warn(\"Your generator is empty.\")\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(10, 1e-3, wd=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:39:15.015122Z",
     "start_time": "2021-01-15T00:39:14.945001Z"
    }
   },
   "outputs": [],
   "source": [
    "test_dl = learn.dls.test_dl(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:39:17.133565Z",
     "start_time": "2021-01-15T00:39:16.287476Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred,_ = learn.get_preds(dl=test_dl)\n",
    "test['Rating'] = np.array(pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-01-15T00:39:26.330439Z",
     "start_time": "2021-01-15T00:39:25.419953Z"
    }
   },
   "outputs": [],
   "source": [
    "test.loc[:,['id','Rating']].to_csv(\"./submission_colab_full.csv\", index=False)"
   ]
  }
 ],
 "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": "3.8.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}