def run_model(model_name, dataset_name, test_size=0.3, clean=False):
print('*' * 70)
print('\tThis is %s model trained on %s with test_size = %.2f' %
(model_name, dataset_name, test_size))
print('*' * 70 + '\n')
model_manager = utils.ModelManager(dataset_name, test_size)
try:
trainset = model_manager.load_model('trainset')
testset = model_manager.load_model('testset')
except OSError:
ratings = DataSet.load_dataset(name=dataset_name)
trainset, testset = DataSet.train_test_split(ratings,
test_size=test_size)
model_manager.save_model(trainset, 'trainset')
model_manager.save_model(testset, 'testset')
'''Do you want to clean workspace and retrain model again?'''
'''if you want to change test_size or retrain model, please set clean_workspace True'''
model_manager.clean_workspace(clean)
if model_name == 'UserCF':
model = UserBasedCF()
elif model_name == 'ItemCF':
model = ItemBasedCF()
elif model_name == 'Random':
model = RandomPredict()
elif model_name == 'MostPopular':
model = MostPopular()
elif model_name == 'UserCF-IIF':
model = UserBasedCF(use_iif_similarity=True)
elif model_name == 'ItemCF-IUF':
model = ItemBasedCF(use_iuf_similarity=True)
else:
raise ValueError('No model named' + model_name)
model.fit(trainset)
recommend_test(model, [1, 100, 233, 666, 888])
model.test(testset)
def fit(self, trainset):
"""
Fit the trainset by calculate movie similarity matrix.
:param trainset: train dataset
:return: None
"""
model_manager = utils.ModelManager()
try:
self.movie_sim_mat = model_manager.load_model(
'movie_sim_mat-iif' if self.
use_iuf_similarity else 'movie_sim_mat')
self.movie_popular = model_manager.load_model('movie_popular')
self.movie_count = model_manager.load_model('movie_count')
self.trainset = model_manager.load_model('trainset')
print(
'Movie similarity model has saved before.\nLoad model success...\n'
)
except OSError:
print('No model saved before.\nTrain a new model...')
self.movie_sim_mat, self.movie_popular, self.movie_count = \
similarity.calculate_item_similarity(trainset=trainset,
use_iuf_similarity=self.use_iuf_similarity)
self.trainset = trainset
print('Train a new model success.')
if self.save_model:
model_manager.save_model(
self.movie_sim_mat, 'movie_sim_mat-iif'
if self.use_iuf_similarity else 'movie_sim_mat')
model_manager.save_model(self.movie_popular, 'movie_popular')
model_manager.save_model(self.movie_count, 'movie_count')
model_manager.save_model(self.trainset, 'trainset')
print('The new model has saved success.\n')
def fit(self, trainset):
"""
Fit the trainset via count movies.
:param trainset: train dataset
:return: None
"""
model_manager = utils.ModelManager()
try:
self.movie_popular = model_manager.load_model('movie_popular')
self.movie_count = model_manager.load_model('movie_count')
self.trainset = model_manager.load_model('trainset')
self.total_movies = model_manager.load_model('total_movies')
print(
'RandomPredict model has saved before.\nLoad model success...\n'
)
except OSError:
print('No model saved before.\nTrain a new model...')
self.trainset = trainset
self.movie_popular, self.movie_count = similarity.calculate_movie_popular(
trainset)
self.total_movies = list(self.movie_popular.keys())
print('Train a new model success.')
if self.save_model:
model_manager.save_model(self.movie_popular, 'movie_popular')
model_manager.save_model(self.movie_count, 'movie_count')
model_manager.save_model(self.total_movies, 'total_movies')
print('The new model has saved success.\n')
def fit(self, trainset):
"""
Fit the trainset by optimize the P and Q.
:param trainset: train dataset
:return: None
"""
self.trainset = trainset
self.users_set, self.items_set, self.items_list, self.item_popular, self.items_count = \
self.init_users_items_set(trainset)
model_manager = utils.ModelManager()
try:
self.P = model_manager.load_model(self.model_name + '-P')
self.Q = model_manager.load_model(self.model_name + '-Q')
print(
'User origin similarity model has saved before.\nLoad model success...\n'
)
except OSError:
print('No model saved before.\nTrain a new model...')
self.init_model(self.users_set, self.items_set, self.K)
self.train(self.trainset)
print('Train a new model success.')
if self.save_model:
model_manager.save_model(self.P, self.model_name + '-P')
model_manager.save_model(self.Q, self.model_name + '-Q')
print('The new model has saved success.\n')
return self.P, self.Q
def runModel(model_name, dataset_name, testSize=0.3, clean=False):
print('*' * 60)
print('This is %s model trained on %s with testSize = %.2f' %
(model_name, dataset_name, testSize))
print('*' * 60 + '\n')
modelManager = utils.ModelManager(dataset_name, testSize)
try:
trainset = modelManager.loadModel('trainset')
testset = modelManager.loadModel('testset')
except OSError:
ratings = DataSet.loadDataset(name=dataset_name)
trainset, testset = DataSet.splitData(ratings, testSize=testSize)
modelManager.saveModel(trainset, 'trainset')
modelManager.saveModel(testset, 'testset')
modelManager.clearWorkspace(clean)
if modelType == 'ItemCF':
model = ItemBasedCF(iufSimilarity=True)
elif modelType == 'UserCF':
model = UserBasedCF(use_iif_similarity=True)
model.fit(trainset)
recommendTest(model, [1, 100, 233, 666])
model.test(trainset)
def fit(self, trainset):
model_manager = utils.ModelManager()
try:
self.user_sim_mat = model_manager.loadModel(
'user_sim_mat-iif' if self.
use_iif_similarity else 'user_sim_mat')
self.movie_popular = model_manager.loadModel('movie_popular')
self.movie_count = model_manager.loadModel('movie_count')
self.trainset = model_manager.loadModel('trainset')
print(
'User origin similarity model has saved before.\nLoad model success...\n'
)
except OSError:
print('No model saved before.\nTrain a new model...')
self.user_sim_mat, self.movie_popular, self.movie_count = \
UserSimilarity.calculate_user_similarity(trainset=trainset,
use_iif_similarity=self.use_iif_similarity)
self.trainset = trainset
print('Train a new model success.')
if self.save_model:
model_manager.saveModel(
self.user_sim_mat, 'user_sim_mat-iif'
if self.use_iif_similarity else 'user_sim_mat')
model_manager.saveModel(self.movie_popular, 'movie_popular')
model_manager.saveModel(self.movie_count, 'movie_count')
print('The new model has saved success.\n')
def fit(self, trainset):
modelManager = utils.ModelManager()
try:
self.simMat = modelManager.loadModel(
'simMat-iif' if self.iufSimilarity else 'simMat')
self.moviePopular = modelManager.loadModel('moviePopular')
self.movieCount = modelManager.loadModel('movieCount')
self.trainset = modelManager.loadModel('trainset')
except OSError:
print(
'No model for the given description saved before.\nTraining a new model...'
)
self.simMat, self.moviePopular, self.movieCount = \
ItemSimilarity.calculateItemSimilarity(trainset=trainset, iufSimilarity=self.iufSimilarity)
self.trainset = trainset
if self.saveModel:
modelManager.saveModel(
self.simMat,
'simMat-iif' if self.iufSimilarity else 'simMat')
modelManager.saveModel(self.moviePopular, 'moviePopular')
modelManager.saveModel(self.movieCount, 'movieCount')
modelManager.saveModel(self.trainset, 'trainset')
