2-4-LDA parameter tuning on validation set.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import datetime\n",
    "import os\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "from src.config import DATA_DIR, MODEL_DIR, PARAM_DIR\n",
    "from src.recommendation.LDA import train_lda_model, model_type\n",
    "from src.recommendation.all_evaluation import _results\n",
    "from src.recommendation.evaluation import evaluate\n",
    "from src.recommendation.mixture_functions import get_train\n",
    "from src.util.data_io import get_dataset\n",
    "from src.util.eval_measures import eval_lst\n",
    "from src.util.io import save_pickle\n",
    "\n",
    "FMT = '%Y-%m-%d %H:%M:%S'\n",
    "train_model, model_type = train_lda_model, model_type\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'lda_model'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_type\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = os.path.join(DATA_DIR, 'day')\n",
    "results_dir = os.path.join(MODEL_DIR, 'tuning')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['recall', [5]], ['precision', [5]], ['nDCG', [5]]]\n"
     ]
    }
   ],
   "source": [
    "print(eval_lst)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# subset Jan 20 - 22, 2015\n",
    "datasets = ['120', '121', '122']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "k_vals = [5, 10, 50, 100, 500]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "overwrite = False\n",
    "save_multinomials = True\n",
    "\n",
    "\n",
    "def hours_minutes_seconds(td):\n",
    "    return '{:02}:{:02}:{:02}'.format(td.seconds // 3600, (td.seconds // 60) % 60, td.seconds % 60)\n",
    "\n",
    "\n",
    "def eval_dataset(dataset, k_vals):\n",
    "    result_dic = {}\n",
    "\n",
    "    for n_components in k_vals:\n",
    "        print(n_components)\n",
    "        train, val, test = get_dataset(dataset, data_dir)\n",
    "        start_time = datetime.datetime.now()\n",
    "        all_multinomials = train_model(train, val, test, results_dir,\n",
    "                                       dataset, overwrite=False, save_multinomials=True,\n",
    "                                       n_components=n_components, is_tune=True)\n",
    "        end_time = datetime.datetime.now()\n",
    "        time_used = hours_minutes_seconds(end_time - start_time)\n",
    "        print('time_used', time_used)\n",
    "        dataset, eval_result = _results(dataset, eval_lst, val, all_multinomials)\n",
    "        eval_result.append(['time_used', time_used])\n",
    "        result_dic.update({n_components: eval_result})\n",
    "\n",
    "    df = pd.DataFrame(result_dic).T\n",
    "    df.columns = [item for sublist in list(result_dic.values())[0] for item in sublist[:1]]\n",
    "    for col in df.columns:\n",
    "        df[col] = df[col].apply(lambda s: s[1])\n",
    "\n",
    "    return df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5\n",
      "2019-10-22 13:23:59,182 INFO --> Loading data for 120\n",
      "time_used 00:00:16\n",
      "recall@5: 0.05069\n",
      "precision@5: 0.05519\n",
      "nDCG@5: 0.07130\n",
      "10\n",
      "2019-10-22 13:25:06,894 INFO --> Loading data for 120\n",
      "time_used 00:00:05\n",
      "recall@5: 0.04832\n",
      "precision@5: 0.05315\n",
      "nDCG@5: 0.06941\n",
      "50\n",
      "2019-10-22 13:26:02,659 INFO --> Loading data for 120\n",
      "time_used 00:00:05\n",
      "recall@5: 0.05439\n",
      "precision@5: 0.05513\n",
      "nDCG@5: 0.07323\n",
      "100\n",
      "2019-10-22 13:26:58,851 INFO --> Loading data for 120\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n",
      "500\n",
      "2019-10-22 13:27:33,323 INFO --> Loading data for 120\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n",
      "5\n",
      "2019-10-22 13:28:08,130 INFO --> Loading data for 121\n",
      "time_used 00:00:16\n",
      "recall@5: 0.04836\n",
      "precision@5: 0.05383\n",
      "nDCG@5: 0.07025\n",
      "10\n",
      "2019-10-22 13:29:19,662 INFO --> Loading data for 121\n",
      "time_used 00:00:05\n",
      "recall@5: 0.04878\n",
      "precision@5: 0.05537\n",
      "nDCG@5: 0.07095\n",
      "50\n",
      "2019-10-22 13:30:20,521 INFO --> Loading data for 121\n",
      "time_used 00:00:05\n",
      "recall@5: 0.05281\n",
      "precision@5: 0.05471\n",
      "nDCG@5: 0.07208\n",
      "100\n",
      "2019-10-22 13:31:21,085 INFO --> Loading data for 121\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n",
      "500\n",
      "2019-10-22 13:31:57,874 INFO --> Loading data for 121\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n",
      "5\n",
      "2019-10-22 13:32:34,869 INFO --> Loading data for 122\n",
      "time_used 00:00:16\n",
      "recall@5: 0.05289\n",
      "precision@5: 0.05865\n",
      "nDCG@5: 0.07578\n",
      "10\n",
      "2019-10-22 13:33:45,364 INFO --> Loading data for 122\n",
      "time_used 00:00:06\n",
      "recall@5: 0.05341\n",
      "precision@5: 0.06093\n",
      "nDCG@5: 0.07739\n",
      "50\n",
      "2019-10-22 13:34:45,640 INFO --> Loading data for 122\n",
      "time_used 00:00:05\n",
      "recall@5: 0.05695\n",
      "precision@5: 0.05959\n",
      "nDCG@5: 0.07815\n",
      "100\n",
      "2019-10-22 13:35:45,657 INFO --> Loading data for 122\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n",
      "500\n",
      "2019-10-22 13:36:21,971 INFO --> Loading data for 122\n",
      "time_used 00:00:03\n",
      "recall@5: 0.00017\n",
      "precision@5: 0.00019\n",
      "nDCG@5: 0.00020\n"
     ]
    }
   ],
   "source": [
    "dfs = dict()\n",
    "for dataset in datasets:\n",
    "    dataset = str(dataset)\n",
    "    df = eval_dataset(dataset, k_vals)\n",
    "    dfs.update({dataset: df})\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>recall@5</th>\n",
       "      <th>precision@5</th>\n",
       "      <th>nDCG@5</th>\n",
       "      <th>time_used</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.052891</td>\n",
       "      <td>0.058647</td>\n",
       "      <td>0.075780</td>\n",
       "      <td>00:00:16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.053406</td>\n",
       "      <td>0.060930</td>\n",
       "      <td>0.077386</td>\n",
       "      <td>00:00:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>0.056951</td>\n",
       "      <td>0.059593</td>\n",
       "      <td>0.078152</td>\n",
       "      <td>00:00:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.000166</td>\n",
       "      <td>0.000194</td>\n",
       "      <td>0.000201</td>\n",
       "      <td>00:00:03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500</th>\n",
       "      <td>0.000166</td>\n",
       "      <td>0.000194</td>\n",
       "      <td>0.000201</td>\n",
       "      <td>00:00:03</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     recall@5  precision@5    nDCG@5 time_used\n",
       "5    0.052891     0.058647  0.075780  00:00:16\n",
       "10   0.053406     0.060930  0.077386  00:00:06\n",
       "50   0.056951     0.059593  0.078152  00:00:05\n",
       "100  0.000166     0.000194  0.000201  00:00:03\n",
       "500  0.000166     0.000194  0.000201  00:00:03"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['recall@5', 'precision@5', 'nDCG@5', 'time_used']"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eval_measure = df.columns.tolist()\n",
    "eval_measure\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_metric(col):\n",
    "    try:\n",
    "        df_plot = pd.DataFrame({c: df[col] for c, df in dfs.items()})\n",
    "        df_plot.plot(logx=True, xticks=k_vals, title=col.upper() + ' for different k')\n",
    "    except:\n",
    "        # time is not numeric value\n",
    "        return df_plot\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "120\n",
      "50\n",
      "121\n",
      "50\n",
      "122\n",
      "50\n"
     ]
    }
   ],
   "source": [
    "for k, df in dfs.items():\n",
    "    print(k)\n",
    "    print(df['nDCG@5'].idxmax())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          120       121       122\n",
      "5    00:00:16  00:00:16  00:00:16\n",
      "10   00:00:05  00:00:05  00:00:06\n",
      "50   00:00:05  00:00:05  00:00:05\n",
      "100  00:00:03  00:00:03  00:00:03\n",
      "500  00:00:03  00:00:03  00:00:03\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for col in eval_measure:\n",
    "    df_time = plot_metric(col)\n",
    "print(df_time)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_pickle(os.path.join(PARAM_DIR, model_type), 50, False)\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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}