final-lightfm.ipynb

{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 5GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","execution_count":1,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"import numpy as np\n\nfrom lightfm.datasets import fetch_movielens\nfrom lightfm import LightFM\nfrom lightfm.evaluation import precision_at_k\nfrom lightfm.evaluation import recall_at_k\nfrom lightfm.evaluation import auc_score\nfrom lightfm.data import Dataset\n\nfrom lightfm.cross_validation import random_train_test_split\nimport scipy.sparse as sp\nfrom scipy.sparse import csr_matrix\nfrom sklearn.model_selection import train_test_split\nimport itertools","execution_count":2,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# movielens dataset with 100k movie ratings from 1k users on 1700 movies\ndata = fetch_movielens(min_rating=2.5)\nprint(data)","execution_count":3,"outputs":[{"output_type":"stream","text":"{'train': <943x1682 sparse matrix of type '<class 'numpy.int32'>'\n\twith 74627 stored elements in COOrdinate format>, 'test': <943x1682 sparse matrix of type '<class 'numpy.int32'>'\n\twith 7893 stored elements in COOrdinate format>, 'item_features': <1682x1682 sparse matrix of type '<class 'numpy.float32'>'\n\twith 1682 stored elements in Compressed Sparse Row format>, 'item_feature_labels': array(['Toy Story (1995)', 'GoldenEye (1995)', 'Four Rooms (1995)', ...,\n       'Sliding Doors (1998)', 'You So Crazy (1994)',\n       'Scream of Stone (Schrei aus Stein) (1991)'], dtype=object), 'item_labels': array(['Toy Story (1995)', 'GoldenEye (1995)', 'Four Rooms (1995)', ...,\n       'Sliding Doors (1998)', 'You So Crazy (1994)',\n       'Scream of Stone (Schrei aus Stein) (1991)'], dtype=object)}\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_set = data['train']\ntest_set = data['test']","execution_count":4,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# create hybrid model, CB+CF\nmodel = LightFM(learning_rate=0.05,loss='warp')\n\n#train model\nmodel.fit(data['train'],epochs=20,num_threads=2)","execution_count":5,"outputs":[{"output_type":"execute_result","execution_count":5,"data":{"text/plain":"<lightfm.lightfm.LightFM at 0x7f2e614c8090>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"result=[]\nk = 10\nuserID_list = [2,10,60]\nknown_positives =[]\ntop_movies = []\ndef lightfm_recommender(model,data,user_ids):\n#     # of users and items usinf shape\n    no_users,no_movies = data['train'].shape\n    \n#     generate recommendations for each user we input\n    for user_id in user_ids:\n#         movies already liked by user so far\n        known_positives = data['item_labels'][data['train'].tocsr() [user_id].indices]\n#         movies our model predicts they will like\n        scores = model.predict(user_id, np.arange(no_movies))\n#      rank them in order of most liked to least\n        top_movies = data['item_labels'][np.argsort(-scores)]\n#       print out the results\n        print(\"User %s\" % user_id)\n#         userID_list.append(user_id)\n        print(\"     Known positives:\")\n        \n        for x in known_positives[:k]:\n            print(\"        %s\" % x)\n            \n        print(\"     Recommended:\")\n        for x in top_movies[:k]:\n            print(\"        %s\" % x)\n        \n        result.append([user_id,known_positives[:k],top_movies[:k]])\n        #result_df = pd.DataFrame(result)\n        #result_df.rename(columns={0:'User', 1:'known Positives',2:'Top Movies Recommended'},inplace=True)\n    return result\n    \n            \nresult = lightfm_recommender(model, data, userID_list) \n\n# result.append([userID_list,known_positives[:k],top_movies[:k]])\nresult_df = pd.DataFrame(result)\nresult_df.rename(columns={0:'User', 1:'known Positives',2:'Top Movies Recommended'},inplace=True)\n","execution_count":10,"outputs":[{"output_type":"stream","text":"User 2\n     Known positives:\n        Return of the Jedi (1983)\n        Event Horizon (1997)\n        Chasing Amy (1997)\n        Starship Troopers (1997)\n        Hoodlum (1997)\n        Ulee's Gold (1997)\n        Devil's Advocate, The (1997)\n        Schindler's List (1993)\n        Paradise Lost: The Child Murders at Robin Hood Hills (1996)\n        Mother (1996)\n     Recommended:\n        Titanic (1997)\n        L.A. Confidential (1997)\n        Air Force One (1997)\n        Contact (1997)\n        Cop Land (1997)\n        Devil's Advocate, The (1997)\n        English Patient, The (1996)\n        G.I. Jane (1997)\n        Seven Years in Tibet (1997)\n        Good Will Hunting (1997)\nUser 10\n     Known positives:\n        Babe (1995)\n        Dead Man Walking (1995)\n        Mr. Holland's Opus (1995)\n        Braveheart (1995)\n        Rumble in the Bronx (1995)\n        Birdcage, The (1996)\n        Apollo 13 (1995)\n        Batman Forever (1995)\n        Strange Days (1995)\n        To Wong Foo, Thanks for Everything! Julie Newmar (1995)\n     Recommended:\n        Star Wars (1977)\n        Return of the Jedi (1983)\n        Raiders of the Lost Ark (1981)\n        Princess Bride, The (1987)\n        Monty Python and the Holy Grail (1974)\n        Back to the Future (1985)\n        Empire Strikes Back, The (1980)\n        Fargo (1996)\n        Indiana Jones and the Last Crusade (1989)\n        Toy Story (1995)\nUser 60\n     Known positives:\n        Contact (1997)\n        Full Monty, The (1997)\n        Air Force One (1997)\n        Rainmaker, The (1997)\n        Game, The (1997)\n        Tomorrow Never Dies (1997)\n     Recommended:\n        Contact (1997)\n        L.A. Confidential (1997)\n        English Patient, The (1996)\n        Air Force One (1997)\n        Titanic (1997)\n        Full Monty, The (1997)\n        Scream (1996)\n        Good Will Hunting (1997)\n        Conspiracy Theory (1997)\n        Game, The (1997)\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"result=[]\nknown_values =[]\nk = 10\nuserID_list = [2,10,60]\nknown_positives =[]\ntop_movies = []\ndef lightfm_recommender(model,data,user_ids):\n#     # of users and items usinf shape\n    no_users,no_movies = data['train'].shape\n    \n#     generate recommendations for each user we input\n    for user_id in user_ids:\n#         movies already liked by user so far\n        known_positives = data['item_labels'][data['train'].tocsr() [user_id].indices]\n#         movies our model predicts they will like\n        scores = model.predict(user_id, np.arange(no_movies))\n#      rank them in order of most liked to least\n        top_movies = data['item_labels'][np.argsort(-scores)]\n#       print out the results\n        print(\"User %s\" % user_id)\n#         userID_list.append(user_id)\n        print(\"     Known positives:\")\n        \n        for x in known_positives[:k]:\n            print(\"        %s\" % x)\n            known_values.append([user_id,x])\n            \n        print(\"     Recommended:\")\n        for x in top_movies[:k]:\n            print(\"        %s\" % x)\n            result.append([user_id,x])\n    return known_values,result\n             \nknown_values, result = lightfm_recommender(model, data, userID_list)\n\nknown_values_df = pd.DataFrame(known_values)\nknown_values_df.rename(columns={0:'User',1:'Known Positives'},inplace=True)\nresult_df = pd.DataFrame(result)\nresult_df.rename(columns={0:'User',1:'Top Movies Recommended'},inplace=True)\n","execution_count":22,"outputs":[{"output_type":"stream","text":"User 2\n     Known positives:\n        Return of the Jedi (1983)\n        Event Horizon (1997)\n        Chasing Amy (1997)\n        Starship Troopers (1997)\n        Hoodlum (1997)\n        Ulee's Gold (1997)\n        Devil's Advocate, The (1997)\n        Schindler's List (1993)\n        Paradise Lost: The Child Murders at Robin Hood Hills (1996)\n        Mother (1996)\n     Recommended:\n        Titanic (1997)\n        L.A. Confidential (1997)\n        Air Force One (1997)\n        Contact (1997)\n        Cop Land (1997)\n        Devil's Advocate, The (1997)\n        English Patient, The (1996)\n        G.I. Jane (1997)\n        Seven Years in Tibet (1997)\n        Good Will Hunting (1997)\nUser 10\n     Known positives:\n        Babe (1995)\n        Dead Man Walking (1995)\n        Mr. Holland's Opus (1995)\n        Braveheart (1995)\n        Rumble in the Bronx (1995)\n        Birdcage, The (1996)\n        Apollo 13 (1995)\n        Batman Forever (1995)\n        Strange Days (1995)\n        To Wong Foo, Thanks for Everything! Julie Newmar (1995)\n     Recommended:\n        Star Wars (1977)\n        Return of the Jedi (1983)\n        Raiders of the Lost Ark (1981)\n        Princess Bride, The (1987)\n        Monty Python and the Holy Grail (1974)\n        Back to the Future (1985)\n        Empire Strikes Back, The (1980)\n        Fargo (1996)\n        Indiana Jones and the Last Crusade (1989)\n        Toy Story (1995)\nUser 60\n     Known positives:\n        Contact (1997)\n        Full Monty, The (1997)\n        Air Force One (1997)\n        Rainmaker, The (1997)\n        Game, The (1997)\n        Tomorrow Never Dies (1997)\n     Recommended:\n        Contact (1997)\n        L.A. Confidential (1997)\n        English Patient, The (1996)\n        Air Force One (1997)\n        Titanic (1997)\n        Full Monty, The (1997)\n        Scream (1996)\n        Good Will Hunting (1997)\n        Conspiracy Theory (1997)\n        Game, The (1997)\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"result_df","execution_count":23,"outputs":[{"output_type":"execute_result","execution_count":23,"data":{"text/plain":"    User                     Top Movies Recommended\n0      2                             Titanic (1997)\n1      2                   L.A. Confidential (1997)\n2      2                       Air Force One (1997)\n3      2                             Contact (1997)\n4      2                            Cop Land (1997)\n5      2               Devil's Advocate, The (1997)\n6      2                English Patient, The (1996)\n7      2                           G.I. Jane (1997)\n8      2                Seven Years in Tibet (1997)\n9      2                   Good Will Hunting (1997)\n10    10                           Star Wars (1977)\n11    10                  Return of the Jedi (1983)\n12    10             Raiders of the Lost Ark (1981)\n13    10                 Princess Bride, The (1987)\n14    10     Monty Python and the Holy Grail (1974)\n15    10                  Back to the Future (1985)\n16    10            Empire Strikes Back, The (1980)\n17    10                               Fargo (1996)\n18    10  Indiana Jones and the Last Crusade (1989)\n19    10                           Toy Story (1995)\n20    60                             Contact (1997)\n21    60                   L.A. Confidential (1997)\n22    60                English Patient, The (1996)\n23    60                       Air Force One (1997)\n24    60                             Titanic (1997)\n25    60                     Full Monty, The (1997)\n26    60                              Scream (1996)\n27    60                   Good Will Hunting (1997)\n28    60                   Conspiracy Theory (1997)\n29    60                           Game, The (1997)","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>User</th>\n      <th>Top Movies Recommended</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>2</td>\n      <td>Titanic (1997)</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>2</td>\n      <td>L.A. Confidential (1997)</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>2</td>\n      <td>Air Force One (1997)</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>2</td>\n      <td>Contact (1997)</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>2</td>\n      <td>Cop Land (1997)</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>2</td>\n      <td>Devil's Advocate, The (1997)</td>\n    </tr>\n    <tr>\n      <th>6</th>\n      <td>2</td>\n      <td>English Patient, The (1996)</td>\n    </tr>\n    <tr>\n      <th>7</th>\n      <td>2</td>\n      <td>G.I. Jane (1997)</td>\n    </tr>\n    <tr>\n      <th>8</th>\n      <td>2</td>\n      <td>Seven Years in Tibet (1997)</td>\n    </tr>\n    <tr>\n      <th>9</th>\n      <td>2</td>\n      <td>Good Will Hunting (1997)</td>\n    </tr>\n    <tr>\n      <th>10</th>\n      <td>10</td>\n      <td>Star Wars (1977)</td>\n    </tr>\n    <tr>\n      <th>11</th>\n      <td>10</td>\n      <td>Return of the Jedi (1983)</td>\n    </tr>\n    <tr>\n      <th>12</th>\n      <td>10</td>\n      <td>Raiders of the Lost Ark (1981)</td>\n    </tr>\n    <tr>\n      <th>13</th>\n      <td>10</td>\n      <td>Princess Bride, The (1987)</td>\n    </tr>\n    <tr>\n      <th>14</th>\n      <td>10</td>\n      <td>Monty Python and the Holy Grail (1974)</td>\n    </tr>\n    <tr>\n      <th>15</th>\n      <td>10</td>\n      <td>Back to the Future (1985)</td>\n    </tr>\n    <tr>\n      <th>16</th>\n      <td>10</td>\n      <td>Empire Strikes Back, The (1980)</td>\n    </tr>\n    <tr>\n      <th>17</th>\n      <td>10</td>\n      <td>Fargo (1996)</td>\n    </tr>\n    <tr>\n      <th>18</th>\n      <td>10</td>\n      <td>Indiana Jones and the Last Crusade (1989)</td>\n    </tr>\n    <tr>\n      <th>19</th>\n      <td>10</td>\n      <td>Toy Story (1995)</td>\n    </tr>\n    <tr>\n      <th>20</th>\n      <td>60</td>\n      <td>Contact (1997)</td>\n    </tr>\n    <tr>\n      <th>21</th>\n      <td>60</td>\n      <td>L.A. Confidential (1997)</td>\n    </tr>\n    <tr>\n      <th>22</th>\n      <td>60</td>\n      <td>English Patient, The (1996)</td>\n    </tr>\n    <tr>\n      <th>23</th>\n      <td>60</td>\n      <td>Air Force One (1997)</td>\n    </tr>\n    <tr>\n      <th>24</th>\n      <td>60</td>\n      <td>Titanic (1997)</td>\n    </tr>\n    <tr>\n      <th>25</th>\n      <td>60</td>\n      <td>Full Monty, The (1997)</td>\n    </tr>\n    <tr>\n      <th>26</th>\n      <td>60</td>\n      <td>Scream (1996)</td>\n    </tr>\n    <tr>\n      <th>27</th>\n      <td>60</td>\n      <td>Good Will Hunting (1997)</td>\n    </tr>\n    <tr>\n      <th>28</th>\n      <td>60</td>\n      <td>Conspiracy Theory (1997)</td>\n    </tr>\n    <tr>\n      <th>29</th>\n      <td>60</td>\n      <td>Game, The (1997)</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"known_values_df","execution_count":24,"outputs":[{"output_type":"execute_result","execution_count":24,"data":{"text/plain":"    User                                    Known Positives\n0      2                          Return of the Jedi (1983)\n1      2                               Event Horizon (1997)\n2      2                                 Chasing Amy (1997)\n3      2                           Starship Troopers (1997)\n4      2                                     Hoodlum (1997)\n5      2                                 Ulee's Gold (1997)\n6      2                       Devil's Advocate, The (1997)\n7      2                            Schindler's List (1993)\n8      2  Paradise Lost: The Child Murders at Robin Hood...\n9      2                                      Mother (1996)\n10    10                                        Babe (1995)\n11    10                            Dead Man Walking (1995)\n12    10                          Mr. Holland's Opus (1995)\n13    10                                  Braveheart (1995)\n14    10                         Rumble in the Bronx (1995)\n15    10                               Birdcage, The (1996)\n16    10                                   Apollo 13 (1995)\n17    10                              Batman Forever (1995)\n18    10                                Strange Days (1995)\n19    10  To Wong Foo, Thanks for Everything! Julie Newm...\n20    60                                     Contact (1997)\n21    60                             Full Monty, The (1997)\n22    60                               Air Force One (1997)\n23    60                              Rainmaker, The (1997)\n24    60                                   Game, The (1997)\n25    60                         Tomorrow Never Dies (1997)","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>User</th>\n      <th>Known Positives</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>2</td>\n      <td>Return of the Jedi (1983)</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>2</td>\n      <td>Event Horizon (1997)</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>2</td>\n      <td>Chasing Amy (1997)</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>2</td>\n      <td>Starship Troopers (1997)</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>2</td>\n      <td>Hoodlum (1997)</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>2</td>\n      <td>Ulee's Gold (1997)</td>\n    </tr>\n    <tr>\n      <th>6</th>\n      <td>2</td>\n      <td>Devil's Advocate, The (1997)</td>\n    </tr>\n    <tr>\n      <th>7</th>\n      <td>2</td>\n      <td>Schindler's List (1993)</td>\n    </tr>\n    <tr>\n      <th>8</th>\n      <td>2</td>\n      <td>Paradise Lost: The Child Murders at Robin Hood...</td>\n    </tr>\n    <tr>\n      <th>9</th>\n      <td>2</td>\n      <td>Mother (1996)</td>\n    </tr>\n    <tr>\n      <th>10</th>\n      <td>10</td>\n      <td>Babe (1995)</td>\n    </tr>\n    <tr>\n      <th>11</th>\n      <td>10</td>\n      <td>Dead Man Walking (1995)</td>\n    </tr>\n    <tr>\n      <th>12</th>\n      <td>10</td>\n      <td>Mr. Holland's Opus (1995)</td>\n    </tr>\n    <tr>\n      <th>13</th>\n      <td>10</td>\n      <td>Braveheart (1995)</td>\n    </tr>\n    <tr>\n      <th>14</th>\n      <td>10</td>\n      <td>Rumble in the Bronx (1995)</td>\n    </tr>\n    <tr>\n      <th>15</th>\n      <td>10</td>\n      <td>Birdcage, The (1996)</td>\n    </tr>\n    <tr>\n      <th>16</th>\n      <td>10</td>\n      <td>Apollo 13 (1995)</td>\n    </tr>\n    <tr>\n      <th>17</th>\n      <td>10</td>\n      <td>Batman Forever (1995)</td>\n    </tr>\n    <tr>\n      <th>18</th>\n      <td>10</td>\n      <td>Strange Days (1995)</td>\n    </tr>\n    <tr>\n      <th>19</th>\n      <td>10</td>\n      <td>To Wong Foo, Thanks for Everything! Julie Newm...</td>\n    </tr>\n    <tr>\n      <th>20</th>\n      <td>60</td>\n      <td>Contact (1997)</td>\n    </tr>\n    <tr>\n      <th>21</th>\n      <td>60</td>\n      <td>Full Monty, The (1997)</td>\n    </tr>\n    <tr>\n      <th>22</th>\n      <td>60</td>\n      <td>Air Force One (1997)</td>\n    </tr>\n    <tr>\n      <th>23</th>\n      <td>60</td>\n      <td>Rainmaker, The (1997)</td>\n    </tr>\n    <tr>\n      <th>24</th>\n      <td>60</td>\n      <td>Game, The (1997)</td>\n    </tr>\n    <tr>\n      <th>25</th>\n      <td>60</td>\n      <td>Tomorrow Never Dies (1997)</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Evaluation\nfrom lightfm.evaluation import precision_at_k\ntrain_p_at_k = precision_at_k(model, data['train'], k=5).mean()\ntest_p_at_k = precision_at_k(model, data['test'], k=5).mean()\nprint(\"PRECISION@K: Train precision: %.4f\" % train_p_at_k)\nprint(\"PRECISION@K: Test precision: %.4f\" % test_p_at_k)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_auc = auc_score(model, train_set).mean()\ntest_auc = auc_score(model, test_set).mean()\nprint('AUC: train %.4f, test %.4f.' % (train_auc, test_auc))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_recall = recall_at_k(model, train_set).mean()\ntest_recall = recall_at_k(model, test_set).mean()\nprint('RECALL@K: train %.4f, test %.4f.' % (train_recall, test_recall))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"dataset = Dataset()\n# dataset.fit(data['item_features'],data['item_feature_labels'])\nmovieXuser = ratings_df.pivot(index='movieId',columns='userId',values='rating').fillna(0)\n# print(movieXuser)\n# convert into Sparse matrix\nMxU_mat = csr_matrix(movieXuser.values)\n# print(data['item_features']) #sparse matrix\n# (interactions, weights) = dataset.build_interactions(data['item_features'])\n\n(train_interactions, test_interactions) = random_train_test_split(interactions=data['item_features'], test_percentage=0.2)\ntrain_interactions, test_interactions = train_test_split(data['item_features'], test_size = 0.2, random_state=20)\n\nprint(train_interactions.shape)\nprint(test_interactions.shape)\nprint(train_set.shape)\nprint(test_set.shape)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Evaluation using hyper parameters"},{"metadata":{},"cell_type":"markdown","source":"AUC Score:"},{"metadata":{"trusted":true},"cell_type":"code","source":"def sample_hyperparameters():\n    \"\"\"\n    Yield possible hyperparameter choices.\n    \"\"\"\n\n    while True:\n        yield {\n            \"no_components\": np.random.randint(16, 64),\n            \"learning_schedule\": np.random.choice([\"adagrad\", \"adadelta\"]),\n            \"loss\": np.random.choice([\"bpr\", \"warp\", \"warp-kos\"]),\n            \"learning_rate\": np.random.exponential(0.05),\n            \"item_alpha\": np.random.exponential(1e-8),\n            \"user_alpha\": np.random.exponential(1e-8),\n            \"max_sampled\": np.random.randint(5, 15),\n            \"num_epochs\": np.random.randint(5, 50),\n            \n        }\n\n\ndef random_search_auc(train, test, num_samples=10, num_threads=1,k=5):\n\n    for hyperparams in itertools.islice(sample_hyperparameters(), num_samples):\n        num_epochs = hyperparams.pop(\"num_epochs\")\n        model = LightFM(**hyperparams)\n        model.fit(train, epochs=num_epochs, num_threads=num_threads)\n        score = auc_score(model, test, train_interactions=train, num_threads=num_threads,check_intersections=False).mean()\n        hyperparams[\"num_epochs\"] = num_epochs\n        yield (score, hyperparams, model)\n\n\nif __name__ == \"__main__\":\n    from lightfm.datasets import fetch_movielens\n    data = fetch_movielens()\n    train = data[\"train\"]\n    test = data[\"test\"]\n    (score, hyperparams, model) = max(random_search_auc(train_set, test_set, num_threads=2), key=lambda x: x[0])\n\n    print(\"Best AUC score {} at {}\".format(score, hyperparams))\n\n    ","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Precision@K:"},{"metadata":{"trusted":true},"cell_type":"code","source":"def random_search_precision_at_k(train, test, num_samples=5, num_threads=2,k=5):\n    \n    for hyperparams in itertools.islice(sample_hyperparameters(), num_samples):\n        num_epochs = hyperparams.pop(\"num_epochs\")\n\n        model = LightFM(**hyperparams)\n        model.fit(train, epochs=num_epochs, num_threads=num_threads)\n\n        p_at_k_score = precision_at_k(model, test_interactions=test, train_interactions=train, num_threads=num_threads,k=k,check_intersections=False,preserve_rows=True).mean()\n        \n        hyperparams[\"num_epochs\"] = num_epochs\n\n        yield (p_at_k_score, hyperparams, model)\n\n\nif __name__ == \"__main__\":\n    from lightfm.datasets import fetch_movielens\n\n    data = fetch_movielens()\n    train = data[\"train\"]\n    test = data[\"test\"]\n\n    (p_at_k_score, hyperparams, model) = max(random_search_precision_at_k(train_set, test_set, num_threads=2,k=5), key=lambda x: x[0])\n\n    print(\"Best Precision@k score {} at {}\".format(p_at_k_score, hyperparams))\n","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Recall@K"},{"metadata":{"trusted":true},"cell_type":"code","source":"def random_search_recall_at_k(train, test, num_samples=10, num_threads=1,k=5):\n    \n\n    for hyperparams in itertools.islice(sample_hyperparameters(), num_samples):\n        num_epochs = hyperparams.pop(\"num_epochs\")\n\n        model = LightFM(**hyperparams)\n        model.fit(train, epochs=num_epochs, num_threads=num_threads)\n\n        r_at_k_score = recall_at_k(model, test, train_interactions=train, num_threads=num_threads, k=k).mean()\n        \n        hyperparams[\"num_epochs\"] = num_epochs\n\n        yield (r_at_k_score, hyperparams, model)\n\n\nif __name__ == \"__main__\":\n    from lightfm.datasets import fetch_movielens\n\n    data = fetch_movielens()\n    train = data[\"train\"]\n    test = data[\"test\"]\n\n    (r_at_k_score, hyperparams, model) = max(random_search_recall_at_k(train_set, test_set, num_threads=2), key=lambda x: x[0])\n\n    print(\"Best Recall@k score {} at {}\".format(r_at_k_score, hyperparams))\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Reference: The above program is referred from the documentation of existing hybrod model."}],"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.7.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":4}