s426220/REK-proj-2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
4cf2994aca
REK-proj-2/project_2_recommender_and_evaluation-Copy1.ipynb

1688 lines
64 KiB
Plaintext
Raw Normal View History

Add recommender with HR10 0.116
2021-06-28 20:18:14 +02:00
{
"cells": [
{
"cell_type": "code",
"execution_count": 17,
"id": "alike-morgan",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The autoreload extension is already loaded. To reload it, use:\n",
" %reload_ext autoreload\n"
]
}
],
"source": [
"%matplotlib inline\n",
"%load_ext autoreload\n",
"%autoreload 2\n",
"\n",
"import numpy as np\n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"import seaborn as sns\n",
"from IPython.display import Markdown, display, HTML\n",
"from collections import defaultdict\n",
"\n",
"import torch\n",
"import torch.nn as nn\n",
"import torch.optim as optim\n",
"from livelossplot import PlotLosses\n",
"\n",
"# Fix the dying kernel problem (only a problem in some installations - you can remove it, if it works without it)\n",
"import os\n",
"os.environ['KMP_DUPLICATE_LIB_OK'] = 'True'"
]
},
{
"cell_type": "markdown",
"id": "blessed-knitting",
"metadata": {},
"source": [
"# Load the dataset for recommenders"
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "victorian-bottom",
"metadata": {},
"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>item_id</th>\n",
" <th>term</th>\n",
" <th>length_of_stay_bucket</th>\n",
" <th>rate_plan</th>\n",
" <th>room_segment</th>\n",
" <th>n_people_bucket</th>\n",
" <th>weekend_stay</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>WinterVacation</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[260-360]</td>\n",
" <td>[5-inf]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>WinterVacation</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>3</td>\n",
" <td>2</td>\n",
" <td>WinterVacation</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[2-2]</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>4</td>\n",
" <td>3</td>\n",
" <td>WinterVacation</td>\n",
" <td>[4-7]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>5</td>\n",
" <td>4</td>\n",
" <td>WinterVacation</td>\n",
" <td>[4-7]</td>\n",
" <td>Standard</td>\n",
" <td>[0-160]</td>\n",
" <td>[2-2]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>6</td>\n",
" <td>5</td>\n",
" <td>Easter</td>\n",
" <td>[4-7]</td>\n",
" <td>Standard</td>\n",
" <td>[260-360]</td>\n",
" <td>[5-inf]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>7</td>\n",
" <td>6</td>\n",
" <td>OffSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[260-360]</td>\n",
" <td>[5-inf]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>8</td>\n",
" <td>7</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[1-1]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>9</td>\n",
" <td>8</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[0-160]</td>\n",
" <td>[1-1]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>8</td>\n",
" <td>7</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[1-1]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>8</td>\n",
" <td>7</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[1-1]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>10</td>\n",
" <td>9</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>11</td>\n",
" <td>9</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>12</td>\n",
" <td>10</td>\n",
" <td>HighSeason</td>\n",
" <td>[8-inf]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>14</td>\n",
" <td>11</td>\n",
" <td>HighSeason</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[0-160]</td>\n",
" <td>[3-4]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"data_path = os.path.join(\"data\", \"hotel_data\")\n",
"\n",
"interactions_df = pd.read_csv(os.path.join(data_path, \"hotel_data_interactions_df.csv\"), index_col=0)\n",
"\n",
"base_item_features = ['term', 'length_of_stay_bucket', 'rate_plan', 'room_segment', 'n_people_bucket', 'weekend_stay']\n",
"\n",
"column_values_dict = {\n",
" 'term': ['WinterVacation', 'Easter', 'OffSeason', 'HighSeason', 'LowSeason', 'MayLongWeekend', 'NewYear', 'Christmas'],\n",
" 'length_of_stay_bucket': ['[0-1]', '[2-3]', '[4-7]', '[8-inf]'],\n",
" 'rate_plan': ['Standard', 'Nonref'],\n",
" 'room_segment': ['[0-160]', '[160-260]', '[260-360]', '[360-500]', '[500-900]'],\n",
" 'n_people_bucket': ['[1-1]', '[2-2]', '[3-4]', '[5-inf]'],\n",
" 'weekend_stay': ['True', 'False']\n",
"}\n",
"\n",
"interactions_df.loc[:, 'term'] = pd.Categorical(\n",
" interactions_df['term'], categories=column_values_dict['term'])\n",
"interactions_df.loc[:, 'length_of_stay_bucket'] = pd.Categorical(\n",
" interactions_df['length_of_stay_bucket'], categories=column_values_dict['length_of_stay_bucket'])\n",
"interactions_df.loc[:, 'rate_plan'] = pd.Categorical(\n",
" interactions_df['rate_plan'], categories=column_values_dict['rate_plan'])\n",
"interactions_df.loc[:, 'room_segment'] = pd.Categorical(\n",
" interactions_df['room_segment'], categories=column_values_dict['room_segment'])\n",
"interactions_df.loc[:, 'n_people_bucket'] = pd.Categorical(\n",
" interactions_df['n_people_bucket'], categories=column_values_dict['n_people_bucket'])\n",
"interactions_df.loc[:, 'weekend_stay'] = interactions_df['weekend_stay'].astype('str')\n",
"interactions_df.loc[:, 'weekend_stay'] = pd.Categorical(\n",
" interactions_df['weekend_stay'], categories=column_values_dict['weekend_stay'])\n",
"\n",
"display(HTML(interactions_df.head(15).to_html()))"
]
},
{
"cell_type": "markdown",
"id": "realistic-third",
"metadata": {},
"source": [
"# (Optional) Prepare numerical user features\n",
"\n",
"The method below is left here for convenience if you want to experiment with content-based user features as an input for your neural network."
]
},
{
"cell_type": "code",
"execution_count": 19,
"id": "variable-jaguar",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['user_term_WinterVacation', 'user_term_Easter', 'user_term_OffSeason', 'user_term_HighSeason', 'user_term_LowSeason', 'user_term_MayLongWeekend', 'user_term_NewYear', 'user_term_Christmas', 'user_length_of_stay_bucket_[0-1]', 'user_length_of_stay_bucket_[2-3]', 'user_length_of_stay_bucket_[4-7]', 'user_length_of_stay_bucket_[8-inf]', 'user_rate_plan_Standard', 'user_rate_plan_Nonref', 'user_room_segment_[0-160]', 'user_room_segment_[160-260]', 'user_room_segment_[260-360]', 'user_room_segment_[360-500]', 'user_room_segment_[500-900]', 'user_n_people_bucket_[1-1]', 'user_n_people_bucket_[2-2]', 'user_n_people_bucket_[3-4]', 'user_n_people_bucket_[5-inf]', 'user_weekend_stay_True', 'user_weekend_stay_False']\n"
]
},
{
"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>user_term_WinterVacation</th>\n",
" <th>user_term_Easter</th>\n",
" <th>user_term_OffSeason</th>\n",
" <th>user_term_HighSeason</th>\n",
" <th>user_term_LowSeason</th>\n",
" <th>user_term_MayLongWeekend</th>\n",
" <th>user_term_NewYear</th>\n",
" <th>user_term_Christmas</th>\n",
" <th>user_length_of_stay_bucket_[0-1]</th>\n",
" <th>user_length_of_stay_bucket_[2-3]</th>\n",
" <th>user_length_of_stay_bucket_[4-7]</th>\n",
" <th>user_length_of_stay_bucket_[8-inf]</th>\n",
" <th>user_rate_plan_Standard</th>\n",
" <th>user_rate_plan_Nonref</th>\n",
" <th>user_room_segment_[0-160]</th>\n",
" <th>user_room_segment_[160-260]</th>\n",
" <th>user_room_segment_[260-360]</th>\n",
" <th>user_room_segment_[360-500]</th>\n",
" <th>user_room_segment_[500-900]</th>\n",
" <th>user_n_people_bucket_[1-1]</th>\n",
" <th>user_n_people_bucket_[2-2]</th>\n",
" <th>user_n_people_bucket_[3-4]</th>\n",
" <th>user_n_people_bucket_[5-inf]</th>\n",
" <th>user_weekend_stay_True</th>\n",
" <th>user_weekend_stay_False</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1</td>\n",
" <td>0.130435</td>\n",
" <td>0.0</td>\n",
" <td>0.652174</td>\n",
" <td>0.086957</td>\n",
" <td>0.130435</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.608696</td>\n",
" <td>0.391304</td>\n",
" <td>0.000000</td>\n",
" <td>0.521739</td>\n",
" <td>0.478261</td>\n",
" <td>0.000000</td>\n",
" <td>0.869565</td>\n",
" <td>0.130435</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.000000</td>\n",
" <td>0.739130</td>\n",
" <td>0.173913</td>\n",
" <td>0.086957</td>\n",
" <td>0.782609</td>\n",
" <td>0.217391</td>\n",
" </tr>\n",
" <tr>\n",
" <th>47</th>\n",
" <td>50</td>\n",
" <td>0.043478</td>\n",
" <td>0.0</td>\n",
" <td>0.434783</td>\n",
" <td>0.304348</td>\n",
" <td>0.217391</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.913043</td>\n",
" <td>0.086957</td>\n",
" <td>0.000000</td>\n",
" <td>0.260870</td>\n",
" <td>0.739130</td>\n",
" <td>0.000000</td>\n",
" <td>0.565217</td>\n",
" <td>0.434783</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.000000</td>\n",
" <td>0.173913</td>\n",
" <td>0.521739</td>\n",
" <td>0.304348</td>\n",
" <td>0.782609</td>\n",
" <td>0.217391</td>\n",
" </tr>\n",
" <tr>\n",
" <th>92</th>\n",
" <td>96</td>\n",
" <td>0.083333</td>\n",
" <td>0.0</td>\n",
" <td>0.708333</td>\n",
" <td>0.125000</td>\n",
" <td>0.041667</td>\n",
" <td>0.041667</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.250000</td>\n",
" <td>0.666667</td>\n",
" <td>0.041667</td>\n",
" <td>0.041667</td>\n",
" <td>0.291667</td>\n",
" <td>0.708333</td>\n",
" <td>0.125000</td>\n",
" <td>0.791667</td>\n",
" <td>0.083333</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.041667</td>\n",
" <td>0.333333</td>\n",
" <td>0.541667</td>\n",
" <td>0.083333</td>\n",
" <td>0.750000</td>\n",
" <td>0.250000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>111</th>\n",
" <td>115</td>\n",
" <td>0.727273</td>\n",
" <td>0.0</td>\n",
" <td>0.272727</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.500000</td>\n",
" <td>0.363636</td>\n",
" <td>0.136364</td>\n",
" <td>0.000000</td>\n",
" <td>1.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.818182</td>\n",
" <td>0.181818</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.818182</td>\n",
" <td>0.090909</td>\n",
" <td>0.045455</td>\n",
" <td>0.045455</td>\n",
" <td>0.363636</td>\n",
" <td>0.636364</td>\n",
" </tr>\n",
" <tr>\n",
" <th>675</th>\n",
" <td>706</td>\n",
" <td>0.091988</td>\n",
" <td>0.0</td>\n",
" <td>0.451039</td>\n",
" <td>0.189911</td>\n",
" <td>0.207715</td>\n",
" <td>0.038576</td>\n",
" <td>0.011869</td>\n",
" <td>0.008902</td>\n",
" <td>0.169139</td>\n",
" <td>0.459941</td>\n",
" <td>0.272997</td>\n",
" <td>0.097923</td>\n",
" <td>0.994065</td>\n",
" <td>0.005935</td>\n",
" <td>0.020772</td>\n",
" <td>0.839763</td>\n",
" <td>0.130564</td>\n",
" <td>0.008902</td>\n",
" <td>0.0</td>\n",
" <td>0.041543</td>\n",
" <td>0.094955</td>\n",
" <td>0.738872</td>\n",
" <td>0.124629</td>\n",
" <td>0.676558</td>\n",
" <td>0.323442</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1699</th>\n",
" <td>1736</td>\n",
" <td>0.034483</td>\n",
" <td>0.0</td>\n",
" <td>0.482759</td>\n",
" <td>0.206897</td>\n",
" <td>0.275862</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.241379</td>\n",
" <td>0.551724</td>\n",
" <td>0.206897</td>\n",
" <td>0.000000</td>\n",
" <td>0.172414</td>\n",
" <td>0.827586</td>\n",
" <td>0.000000</td>\n",
" <td>0.931034</td>\n",
" <td>0.068966</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.379310</td>\n",
" <td>0.413793</td>\n",
" <td>0.206897</td>\n",
" <td>0.000000</td>\n",
" <td>0.448276</td>\n",
" <td>0.551724</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7639</th>\n",
" <td>7779</td>\n",
" <td>0.037037</td>\n",
" <td>0.0</td>\n",
" <td>0.296296</td>\n",
" <td>0.259259</td>\n",
" <td>0.370370</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.037037</td>\n",
" <td>0.111111</td>\n",
" <td>0.296296</td>\n",
" <td>0.481481</td>\n",
" <td>0.111111</td>\n",
" <td>1.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.000000</td>\n",
" <td>0.814815</td>\n",
" <td>0.185185</td>\n",
" <td>0.000000</td>\n",
" <td>0.0</td>\n",
" <td>0.000000</td>\n",
" <td>0.037037</td>\n",
" <td>0.740741</td>\n",
" <td>0.222222</td>\n",
" <td>0.814815</td>\n",
" <td>0.185185</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"def n_to_p(l):\n",
" n = sum(l)\n",
" return [x / n for x in l] if n > 0 else l\n",
"\n",
"def calculate_p(x, values):\n",
" counts = [0]*len(values)\n",
" for v in x:\n",
" counts[values.index(v)] += 1\n",
"\n",
" return n_to_p(counts)\n",
"\n",
"def prepare_users_df(interactions_df):\n",
"\n",
" users_df = interactions_df.loc[:, [\"user_id\"]]\n",
" users_df = users_df.groupby(\"user_id\").first().reset_index(drop=False)\n",
" \n",
" user_features = []\n",
"\n",
" for column in base_item_features:\n",
"\n",
" column_values = column_values_dict[column]\n",
" df = interactions_df.loc[:, ['user_id', column]]\n",
" df = df.groupby('user_id').aggregate(lambda x: list(x)).reset_index(drop=False)\n",
"\n",
" def calc_p(x):\n",
" return calculate_p(x, column_values)\n",
"\n",
" df.loc[:, column] = df[column].apply(lambda x: calc_p(x))\n",
"\n",
" p_columns = []\n",
" for i in range(len(column_values)):\n",
" p_columns.append(\"user_\" + column + \"_\" + column_values[i])\n",
" df.loc[:, p_columns[i]] = df[column].apply(lambda x: x[i])\n",
" user_features.append(p_columns[i])\n",
"\n",
" users_df = pd.merge(users_df, df.loc[:, ['user_id'] + p_columns], on=[\"user_id\"])\n",
" \n",
" return users_df, user_features\n",
" \n",
"\n",
"users_df, user_features = prepare_users_df(interactions_df)\n",
"\n",
"print(user_features)\n",
"\n",
"display(HTML(users_df.loc[users_df['user_id'].isin([706, 1736, 7779, 96, 1, 50, 115])].head(15).to_html()))"
]
},
{
"cell_type": "markdown",
"id": "amino-keyboard",
"metadata": {},
"source": [
"# (Optional) Prepare numerical item features\n",
"\n",
"The method below is left here for convenience if you want to experiment with content-based item features as an input for your neural network."
]
},
{
"cell_type": "code",
"execution_count": 20,
"id": "formal-munich",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['term_WinterVacation', 'term_Easter', 'term_OffSeason', 'term_HighSeason', 'term_LowSeason', 'term_MayLongWeekend', 'term_NewYear', 'term_Christmas', 'length_of_stay_bucket_[0-1]', 'length_of_stay_bucket_[2-3]', 'length_of_stay_bucket_[4-7]', 'length_of_stay_bucket_[8-inf]', 'rate_plan_Standard', 'rate_plan_Nonref', 'room_segment_[0-160]', 'room_segment_[160-260]', 'room_segment_[260-360]', 'room_segment_[360-500]', 'room_segment_[500-900]', 'n_people_bucket_[1-1]', 'n_people_bucket_[2-2]', 'n_people_bucket_[3-4]', 'n_people_bucket_[5-inf]', 'weekend_stay_True', 'weekend_stay_False']\n"
]
},
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>item_id</th>\n",
" <th>term_WinterVacation</th>\n",
" <th>term_Easter</th>\n",
" <th>term_OffSeason</th>\n",
" <th>term_HighSeason</th>\n",
" <th>term_LowSeason</th>\n",
" <th>term_MayLongWeekend</th>\n",
" <th>term_NewYear</th>\n",
" <th>term_Christmas</th>\n",
" <th>length_of_stay_bucket_[0-1]</th>\n",
" <th>length_of_stay_bucket_[2-3]</th>\n",
" <th>length_of_stay_bucket_[4-7]</th>\n",
" <th>length_of_stay_bucket_[8-inf]</th>\n",
" <th>rate_plan_Standard</th>\n",
" <th>rate_plan_Nonref</th>\n",
" <th>room_segment_[0-160]</th>\n",
" <th>room_segment_[160-260]</th>\n",
" <th>room_segment_[260-360]</th>\n",
" <th>room_segment_[360-500]</th>\n",
" <th>room_segment_[500-900]</th>\n",
" <th>n_people_bucket_[1-1]</th>\n",
" <th>n_people_bucket_[2-2]</th>\n",
" <th>n_people_bucket_[3-4]</th>\n",
" <th>n_people_bucket_[5-inf]</th>\n",
" <th>weekend_stay_True</th>\n",
" <th>weekend_stay_False</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>3</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>4</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>5</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>6</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"def map_items_to_onehot(df):\n",
" one_hot = pd.get_dummies(df.loc[:, base_item_features])\n",
" df = df.drop(base_item_features, axis = 1)\n",
" df = df.join(one_hot)\n",
" \n",
" return df, list(one_hot.columns)\n",
"\n",
"def prepare_items_df(interactions_df):\n",
" items_df = interactions_df.loc[:, [\"item_id\"] + base_item_features].drop_duplicates()\n",
" \n",
" items_df, item_features = map_items_to_onehot(items_df)\n",
" \n",
" return items_df, item_features\n",
"\n",
"\n",
"items_df, item_features = prepare_items_df(interactions_df)\n",
"\n",
"print(item_features)\n",
"\n",
"display(HTML(items_df.loc[items_df['item_id'].isin([0, 1, 2, 3, 4, 5, 6])].head(15).to_html()))"
]
},
{
"cell_type": "markdown",
"id": "figured-imaging",
"metadata": {},
"source": [
"# Neural network recommender\n",
"\n",
"<span style=\"color:red\"><font size=\"4\">**Task:**</font></span><br> \n",
"Code a recommender based on a neural network model. You are free to choose any network architecture you find appropriate. The network can use the interaction vectors for users and items, embeddings of users and items, as well as user and item features (you can use the features you developed in the first project).\n",
"\n",
"Remember to keep control over randomness - in the init method add the seed as a parameter and initialize the random seed generator with that seed (both for numpy and pytorch):\n",
"\n",
"```python\n",
"self.seed = seed\n",
"self.rng = np.random.RandomState(seed=seed)\n",
"```\n",
"in the network model:\n",
"```python\n",
"self.seed = torch.manual_seed(seed)\n",
"```\n",
"\n",
"You are encouraged to experiment with:\n",
" - the number of layers in the network, the number of neurons and different activation functions,\n",
" - different optimizers and their parameters,\n",
" - batch size and the number of epochs,\n",
" - embedding layers,\n",
" - content-based features of both users and items."
]
},
{
"cell_type": "code",
"execution_count": 21,
"id": "unlike-recipient",
"metadata": {},
"outputs": [],
"source": [
"from recommenders.recommender import Recommender\n",
"\n",
"\n",
"class Net(nn.Module):\n",
" def __init__(self, features_len, output_len):\n",
" super(Net, self).__init__()\n",
" \n",
" print(\"IN:\", features_len, \"OUT:\", output_len)\n",
" \n",
" self.fc1 = nn.Linear(features_len, 150)\n",
" self.fc2 = nn.Linear(150, 50)\n",
" self.fc3 = nn.Linear(50, 25)\n",
" self.fc4 = nn.Linear(25, output_len+500)\n",
" \n",
" def forward(self, x):\n",
" x = F.relu(self.fc1(x))\n",
" x = F.relu(self.fc2(x))\n",
" x = F.relu(self.fc3(x))\n",
" return self.fc4(x)\n",
"\n",
"# class Net(nn.Module):\n",
"# def __init__(self, features_len):\n",
"# super(Net, self).__init__()\n",
"# self.hid1 = nn.Linear(features_len, 10)\n",
"# self.hid2 = nn.Linear(10, 10)\n",
"# self.oupt = nn.Linear(10, 1)\n",
"\n",
"# nn.init.xavier_uniform_(self.hid1.weight)\n",
"# nn.init.zeros_(self.hid1.bias)\n",
"# nn.init.xavier_uniform_(self.hid2.weight)\n",
"# nn.init.zeros_(self.hid2.bias)\n",
"# nn.init.xavier_uniform_(self.oupt.weight)\n",
"# nn.init.zeros_(self.oupt.bias)\n",
"\n",
"# def forward(self, x):\n",
"# z = torch.tanh(self.hid1(x))\n",
"# z = torch.tanh(self.hid2(z))\n",
"# z = torch.sigmoid(self.oupt(z))\n",
"# return z\n",
" \n",
" \n",
"class NNRecommender(Recommender):\n",
" \"\"\"\n",
" Linear recommender class based on user and item features.\n",
" \"\"\"\n",
" \n",
" def generate_negative_interaction(self):\n",
" user_ids = interactions_df['user_id']\n",
" item_ids = interactions_df['item_id']\n",
" \n",
" user_id = user_ids.sample().item()\n",
" item_id = item_ids.sample().item()\n",
" positive_interactions = interactions_df.loc[\n",
" (interactions_df['item_id'] == item_id) & (interactions_df['user_id'] == user_id)]\n",
" \n",
" while not positive_interactions.empty:\n",
" user_id = user_ids.sample().item()\n",
" item_id = item_ids.sample().item()\n",
" positive_interactions = interactions_df.loc[\n",
" (interactions_df['item_id'] == item_id) & (interactions_df['user_id'] == user_id)]\n",
" \n",
" return (user_id, item_id, 0)\n",
" \n",
" def generate_negative_interactions(self, n, interactions_df, cross_df):\n",
" combined_dfs = pd.concat([cross_df, interactions_df[['user_id', 'item_id']]])\n",
" return combined_dfs.drop_duplicates(keep=False).sample(n=n)\n",
" \n",
" \n",
" def __init__(self, seed=6789, n_neg_per_pos=5):\n",
" \"\"\"\n",
" Initialize base recommender params and variables.\n",
" \"\"\"\n",
" self.model = None\n",
" self.n_neg_per_pos = n_neg_per_pos\n",
" \n",
" self.recommender_df = pd.DataFrame(columns=['user_id', 'item_id', 'score'])\n",
" self.users_df = None\n",
" self.user_features = None\n",
" \n",
" self.seed = seed\n",
" self.rng = np.random.RandomState(seed=seed)\n",
" \n",
" def calculate_accuracy(self, y_true, y_pred):\n",
" predicted = y_pred.ge(.5).view(-1)\n",
" return (y_true == predicted).sum().float() / len(y_true)\n",
" \n",
" def round_tensor(self, t, decimal_places=3):\n",
" return round(t.item(), decimal_places)\n",
" \n",
" def fit(self, interactions_df, users_df, items_df):\n",
" \"\"\"\n",
" Training of the recommender.\n",
" \n",
" :param pd.DataFrame interactions_df: DataFrame with recorded interactions between users and items \n",
" defined by user_id, item_id and features of the interaction.\n",
" :param pd.DataFrame users_df: DataFrame with users and their features defined by user_id and the user feature columns.\n",
" :param pd.DataFrame items_df: DataFrame with items and their features defined by item_id and the item feature columns.\n",
" \"\"\"\n",
" \n",
" interactions_df = interactions_df.copy()\n",
" # Prepare users_df and items_df \n",
" # (optional - use only if you want to train a hybrid model with content-based features)\n",
" \n",
" users_df, user_features = prepare_users_df(interactions_df)\n",
" \n",
" self.users_df = users_df\n",
" self.user_features = user_features\n",
" \n",
" items_df, item_features = prepare_items_df(interactions_df)\n",
" items_df = items_df.loc[:, ['item_id'] + item_features]\n",
" \n",
" n_epochs = 51\n",
"\n",
" X = items_df[['term_WinterVacation', 'term_Easter', 'term_OffSeason', 'term_HighSeason', 'term_LowSeason', 'term_MayLongWeekend', 'term_NewYear', 'term_Christmas', 'rate_plan_Standard', 'rate_plan_Nonref', 'room_segment_[0-160]', 'room_segment_[160-260]', 'room_segment_[260-360]', 'room_segment_[360-500]', 'room_segment_[500-900]', 'n_people_bucket_[1-1]', 'n_people_bucket_[2-2]', 'n_people_bucket_[3-4]', 'n_people_bucket_[5-inf]', 'weekend_stay_True', 'weekend_stay_False']]\n",
" y = items_df[['item_id']]\n",
" X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=self.seed)\n",
" \n",
" X_train = torch.from_numpy(X_train.to_numpy()).float()\n",
" y_train = torch.squeeze(torch.from_numpy(y_train.to_numpy()).long())\n",
" X_test = torch.from_numpy(X_test.to_numpy()).float()\n",
" y_test = torch.squeeze(torch.from_numpy(y_test.to_numpy()).long())\n",
" \n",
" self.net = Net(X_train.shape[1], items_df['item_id'].unique().size)\n",
" \n",
" optimizer = optim.Adam(self.net.parameters(), lr=0.05)\n",
" criterion = nn.CrossEntropyLoss()\n",
" \n",
" for epoch in range(n_epochs):\n",
" y_pred = self.net(X_train)\n",
" y_pred = torch.squeeze(y_pred)\n",
" train_loss = criterion(y_pred, y_train)\n",
" \n",
"# if epoch % 5000 == 0:\n",
"# train_acc = self.calculate_accuracy(y_train, y_pred)\n",
"# y_test_pred = self.net(X_test)\n",
"# y_test_pred = torch.squeeze(y_test_pred)\n",
"# test_loss = criterion(y_test_pred, y_test)\n",
"# test_acc = self.calculate_accuracy(y_test, y_test_pred)\n",
"# print(\n",
"# f'''epoch {epoch}\n",
"# Train set - loss: {self.round_tensor(train_loss)}, accuracy: {self.round_tensor(train_acc)}\n",
"# Test set - loss: {self.round_tensor(test_loss)}, accuracy: {self.round_tensor(test_acc)}\n",
"# ''')\n",
" \n",
" optimizer.zero_grad()\n",
" train_loss.backward()\n",
" optimizer.step()\n",
" \n",
" def recommend(self, users_df, items_df, n_recommendations=1):\n",
" \"\"\"\n",
" Serving of recommendations. Scores items in items_df for each user in users_df and returns \n",
" top n_recommendations for each user.\n",
" \n",
" :param pd.DataFrame users_df: DataFrame with users and their features for which recommendations should be generated.\n",
" :param pd.DataFrame items_df: DataFrame with items and their features which should be scored.\n",
" :param int n_recommendations: Number of recommendations to be returned for each user.\n",
" :return: DataFrame with user_id, item_id and score as columns returning n_recommendations top recommendations \n",
" for each user.\n",
" :rtype: pd.DataFrame\n",
" \"\"\"\n",
" \n",
" # Clean previous recommendations (iloc could be used alternatively)\n",
" self.recommender_df = self.recommender_df[:0]\n",
" \n",
" # Prepare users_df and items_df\n",
" # (optional - use only if you want to train a hybrid model with content-based features)\n",
" \n",
" users_df = users_df.loc[:, 'user_id']\n",
" users_df = pd.merge(users_df, self.users_df, on=['user_id'], how='left').fillna(0)\n",
" \n",
" # items_df, item_features = prepare_items_df(items_df)\n",
" # items_df = items_df.loc[:, ['item_id'] + item_features]\n",
" \n",
" # Score the items\n",
" \n",
" recommendations = pd.DataFrame(columns=['user_id', 'item_id', 'score'])\n",
" \n",
" for ix, user in users_df.iterrows():\n",
" prep_user = torch.from_numpy(user[['user_term_WinterVacation', 'user_term_Easter', 'user_term_OffSeason', 'user_term_HighSeason', 'user_term_LowSeason', 'user_term_MayLongWeekend', 'user_term_NewYear', 'user_term_Christmas', 'user_rate_plan_Standard', 'user_rate_plan_Nonref', 'user_room_segment_[0-160]', 'user_room_segment_[160-260]', 'user_room_segment_[260-360]', 'user_room_segment_[360-500]', 'user_room_segment_[500-900]', 'user_n_people_bucket_[1-1]', 'user_n_people_bucket_[2-2]', 'user_n_people_bucket_[3-4]', 'user_n_people_bucket_[5-inf]', 'user_weekend_stay_True', 'user_weekend_stay_False']].to_numpy()).float()\n",
" \n",
" scores = self.net(prep_user).detach().numpy()\n",
" \n",
" chosen_ids = np.argsort(-scores)[:n_recommendations]\n",
" \n",
" recommendations = []\n",
" for item_id in chosen_ids:\n",
" recommendations.append(\n",
" {\n",
" 'user_id': user['user_id'],\n",
" 'item_id': item_id,\n",
" 'score': scores[item_id]\n",
" }\n",
" )\n",
" \n",
" user_recommendations = pd.DataFrame(recommendations)\n",
" \n",
" self.recommender_df = pd.concat([self.recommender_df, user_recommendations])\n",
" \n",
" return self.recommender_df"
]
},
{
"cell_type": "markdown",
"id": "copyrighted-relative",
"metadata": {},
"source": [
"# Quick test of the recommender"
]
},
{
"cell_type": "code",
"execution_count": 22,
"id": "greatest-canon",
"metadata": {},
"outputs": [],
"source": [
"items_df = interactions_df.loc[:, ['item_id'] + base_item_features].drop_duplicates()"
]
},
{
"cell_type": "code",
"execution_count": 23,
"id": "initial-capital",
"metadata": {},
"outputs": [
{
"ename": "NameError",
"evalue": "name 'train_test_split' is not defined",
"output_type": "error",
"traceback": [
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[0;31mNameError\u001b[0m Traceback (most recent call last)",
"\u001b[0;32m<ipython-input-23-851d3aa5378e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m 1\u001b[0m \u001b[0;31m# Fit method\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 2\u001b[0m \u001b[0mnn_recommender\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mNNRecommender\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0mnn_recommender\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minteractions_df\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1000\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 4\u001b[0m \u001b[0;31m# nn_recommender.fit(interactions_df, None, None)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m<ipython-input-21-ca3049874457>\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, interactions_df, users_df, items_df)\u001b[0m\n\u001b[1;32m 114\u001b[0m \u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitems_df\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'term_WinterVacation'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_Easter'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_OffSeason'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_HighSeason'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_LowSeason'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_MayLongWeekend'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_NewYear'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'term_Christmas'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'rate_plan_Standard'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'rate_plan_Nonref'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'room_segment_[0-160]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'room_segment_[160-260]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'room_segment_[260-360]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'room_segment_[360-500]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'room_segment_[500-900]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'n_people_bucket_[1-1]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'n_people_bucket_[2-2]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'n_people_bucket_[3-4]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'n_people_bucket_[5-inf]'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'weekend_stay_True'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'weekend_stay_False'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 115\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitems_df\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'item_id'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 116\u001b[0;31m \u001b[0mX_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtrain_test_split\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrandom_state\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseed\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 117\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 118\u001b[0m \u001b[0mX_train\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_numpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto_numpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;31mNameError\u001b[0m: name 'train_test_split' is not defined"
]
}
],
"source": [
"# Fit method\n",
"nn_recommender = NNRecommender()\n",
"nn_recommender.fit(interactions_df.head(1000), None, None)\n",
"# nn_recommender.fit(interactions_df, None, None)"
]
},
{
"cell_type": "code",
"execution_count": 193,
"id": "digital-consolidation",
"metadata": {
"scrolled": true
},
"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>item_id</th>\n",
" <th>score</th>\n",
" <th>term</th>\n",
" <th>length_of_stay_bucket</th>\n",
" <th>rate_plan</th>\n",
" <th>room_segment</th>\n",
" <th>n_people_bucket</th>\n",
" <th>weekend_stay</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1.0</td>\n",
" <td>88</td>\n",
" <td>37.715969</td>\n",
" <td>WinterVacation</td>\n",
" <td>[0-1]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[2-2]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1.0</td>\n",
" <td>57</td>\n",
" <td>36.182877</td>\n",
" <td>WinterVacation</td>\n",
" <td>[2-3]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[2-2]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1.0</td>\n",
" <td>69</td>\n",
" <td>35.771114</td>\n",
" <td>WinterVacation</td>\n",
" <td>[4-7]</td>\n",
" <td>Standard</td>\n",
" <td>[160-260]</td>\n",
" <td>[2-2]</td>\n",
" <td>True</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Recommender method\n",
"\n",
"recommendations = nn_recommender.recommend(pd.DataFrame([[1]], columns=['user_id']), items_df, 3)\n",
"\n",
"recommendations = pd.merge(recommendations, items_df, on='item_id', how='left')\n",
"display(HTML(recommendations.to_html()))"
]
},
{
"cell_type": "markdown",
"id": "advanced-eleven",
"metadata": {},
"source": [
"# Tuning method"
]
},
{
"cell_type": "code",
"execution_count": 194,
"id": "strange-alaska",
"metadata": {},
"outputs": [],
"source": [
"from evaluation_and_testing.testing import evaluate_train_test_split_implicit\n",
"\n",
"seed = 6789"
]
},
{
"cell_type": "code",
"execution_count": 195,
"id": "stable-theta",
"metadata": {},
"outputs": [],
"source": [
"from hyperopt import hp, fmin, tpe, Trials\n",
"import traceback\n",
"\n",
"def tune_recommender(recommender_class, interactions_df, items_df, \n",
" param_space, max_evals=1, show_progressbar=True, seed=6789):\n",
" # Split into train_validation and test sets\n",
"\n",
" shuffle = np.arange(len(interactions_df))\n",
" rng = np.random.RandomState(seed=seed)\n",
" rng.shuffle(shuffle)\n",
" shuffle = list(shuffle)\n",
"\n",
" train_test_split = 0.8\n",
" split_index = int(len(interactions_df) * train_test_split)\n",
"\n",
" train_validation = interactions_df.iloc[shuffle[:split_index]]\n",
" test = interactions_df.iloc[shuffle[split_index:]]\n",
"\n",
" # Tune\n",
"\n",
" def loss(tuned_params):\n",
" recommender = recommender_class(seed=seed, **tuned_params)\n",
" hr1, hr3, hr5, hr10, ndcg1, ndcg3, ndcg5, ndcg10 = evaluate_train_test_split_implicit(\n",
" recommender, train_validation, items_df, seed=seed)\n",
" return -hr10\n",
"\n",
" n_tries = 1\n",
" succeded = False\n",
" try_id = 0\n",
" while not succeded and try_id < n_tries:\n",
" try:\n",
" trials = Trials()\n",
" best_param_set = fmin(loss, space=param_space, algo=tpe.suggest, \n",
" max_evals=max_evals, show_progressbar=show_progressbar, trials=trials, verbose=True)\n",
" succeded = True\n",
" except:\n",
" traceback.print_exc()\n",
" try_id += 1\n",
" \n",
" if not succeded:\n",
" return None\n",
" \n",
" # Validate\n",
" \n",
" recommender = recommender_class(seed=seed, **best_param_set)\n",
"\n",
" results = [[recommender_class.__name__] + list(evaluate_train_test_split_implicit(\n",
" recommender, {'train': train_validation, 'test': test}, items_df, seed=seed))]\n",
"\n",
" results = pd.DataFrame(results, \n",
" columns=['Recommender', 'HR@1', 'HR@3', 'HR@5', 'HR@10', 'NDCG@1', 'NDCG@3', 'NDCG@5', 'NDCG@10'])\n",
"\n",
" display(HTML(results.to_html()))\n",
" \n",
" return best_param_set"
]
},
{
"cell_type": "markdown",
"id": "reliable-switzerland",
"metadata": {},
"source": [
"## Tuning of the recommender\n",
"\n",
"<span style=\"color:red\"><font size=\"4\">**Task:**</font></span><br> \n",
"Tune your model using the code below. You only need to put the class name of your recommender and choose an appropriate parameter space."
]
},
{
"cell_type": "code",
"execution_count": 196,
"id": "obvious-astrology",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"IN: \n",
"21 \n",
"OUT: \n",
"691 \n",
"100%|██████████| 10/10 [18:34<00:00, 111.50s/trial, best loss: -0.04424416222859484]\n",
"IN: 21 OUT: 736\n"
]
},
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Recommender</th>\n",
" <th>HR@1</th>\n",
" <th>HR@3</th>\n",
" <th>HR@5</th>\n",
" <th>HR@10</th>\n",
" <th>NDCG@1</th>\n",
" <th>NDCG@3</th>\n",
" <th>NDCG@5</th>\n",
" <th>NDCG@10</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>NNRecommender</td>\n",
" <td>0.010201</td>\n",
" <td>0.020072</td>\n",
" <td>0.026324</td>\n",
" <td>0.035538</td>\n",
" <td>0.010201</td>\n",
" <td>0.01574</td>\n",
" <td>0.018216</td>\n",
" <td>0.021141</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Best parameters:\n",
"{'n_neg_per_pos': 9.0}\n"
]
}
],
"source": [
"param_space = {\n",
" 'n_neg_per_pos': hp.quniform('n_neg_per_pos', 1, 10, 1)\n",
"}\n",
"items_df['item_id'].unique().size\n",
"\n",
"best_param_set = tune_recommender(NNRecommender, interactions_df, items_df,\n",
" param_space, max_evals=10, show_progressbar=True, seed=seed)\n",
"\n",
"print(\"Best parameters:\")\n",
"print(best_param_set)"
]
},
{
"cell_type": "markdown",
"id": "accredited-strap",
"metadata": {},
"source": [
"# Final evaluation\n",
"\n",
"<span style=\"color:red\"><font size=\"4\">**Task:**</font></span><br> \n",
"Run the final evaluation of your recommender and present its results against the Amazon and Netflix recommenders' results. You just need to give the class name of your recommender and its tuned parameters below."
]
},
{
"cell_type": "code",
"execution_count": 198,
"id": "given-homework",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"IN: 21 OUT: 736\n"
]
},
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Recommender</th>\n",
" <th>HR@1</th>\n",
" <th>HR@3</th>\n",
" <th>HR@5</th>\n",
" <th>HR@10</th>\n",
" <th>NDCG@1</th>\n",
" <th>NDCG@3</th>\n",
" <th>NDCG@5</th>\n",
" <th>NDCG@10</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>NNRecommender</td>\n",
" <td>0.003949</td>\n",
" <td>0.015137</td>\n",
" <td>0.019743</td>\n",
" <td>0.026654</td>\n",
" <td>0.003949</td>\n",
" <td>0.010361</td>\n",
" <td>0.01223</td>\n",
" <td>0.014409</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"nn_recommender = NNRecommender(n_neg_per_pos=9) # Initialize your recommender here\n",
"\n",
"# Give the name of your recommender in the line below\n",
"nn_tts_results = [['NNRecommender'] + list(evaluate_train_test_split_implicit(\n",
" nn_recommender, interactions_df, items_df))]\n",
"\n",
"nn_tts_results = pd.DataFrame(\n",
" nn_tts_results, columns=['Recommender', 'HR@1', 'HR@3', 'HR@5', 'HR@10', 'NDCG@1', 'NDCG@3', 'NDCG@5', 'NDCG@10'])\n",
"\n",
"display(HTML(nn_tts_results.to_html()))"
]
},
{
"cell_type": "code",
"execution_count": 199,
"id": "suited-nomination",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Recommender</th>\n",
" <th>HR@1</th>\n",
" <th>HR@3</th>\n",
" <th>HR@5</th>\n",
" <th>HR@10</th>\n",
" <th>NDCG@1</th>\n",
" <th>NDCG@3</th>\n",
" <th>NDCG@5</th>\n",
" <th>NDCG@10</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>AmazonRecommender</td>\n",
" <td>0.042119</td>\n",
" <td>0.10464</td>\n",
" <td>0.140507</td>\n",
" <td>0.199408</td>\n",
" <td>0.042119</td>\n",
" <td>0.076826</td>\n",
" <td>0.091797</td>\n",
" <td>0.110705</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from recommenders.amazon_recommender import AmazonRecommender\n",
"\n",
"amazon_recommender = AmazonRecommender()\n",
"\n",
"amazon_tts_results = [['AmazonRecommender'] + list(evaluate_train_test_split_implicit(\n",
" amazon_recommender, interactions_df, items_df))]\n",
"\n",
"amazon_tts_results = pd.DataFrame(\n",
" amazon_tts_results, columns=['Recommender', 'HR@1', 'HR@3', 'HR@5', 'HR@10', 'NDCG@1', 'NDCG@3', 'NDCG@5', 'NDCG@10'])\n",
"\n",
"display(HTML(amazon_tts_results.to_html()))"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "conservative-remedy",
"metadata": {},
"outputs": [],
"source": [
"from recommenders.netflix_recommender import NetflixRecommender\n",
"\n",
"netflix_recommender = NetflixRecommender(n_epochs=30, print_type='live')\n",
"\n",
"netflix_tts_results = [['NetflixRecommender'] + list(evaluate_train_test_split_implicit(\n",
" netflix_recommender, interactions_df, items_df))]\n",
"\n",
"netflix_tts_results = pd.DataFrame(\n",
" netflix_tts_results, columns=['Recommender', 'HR@1', 'HR@3', 'HR@5', 'HR@10', 'NDCG@1', 'NDCG@3', 'NDCG@5', 'NDCG@10'])\n",
"\n",
"display(HTML(netflix_tts_results.to_html()))"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "moderate-printing",
"metadata": {},
"outputs": [],
"source": [
"tts_results = pd.concat([nn_tts_results, amazon_tts_results, netflix_tts_results]).reset_index(drop=True)\n",
"display(HTML(tts_results.to_html()))"
]
},
{
"cell_type": "markdown",
"id": "uniform-vegetable",
"metadata": {},
"source": [
"# Summary\n",
"\n",
"<span style=\"color:red\"><font size=\"4\">**Task:**</font></span><br> \n",
"Write a summary of your experiments. What worked well and what did not? What are your thoughts how could you possibly further improve the model?"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "declared-howard",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "rek_uno",
"language": "python",
"name": "rek_uno"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.8"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Reference in New Issue Copy Permalink