def shape(self):
nb_users = global_parameters(self.sets_parameters)[0]
nb_movies = global_parameters(self.sets_parameters)[1]
if self.sets_parameters['learning_type'] == 'V':
nb_users, nb_movies = permute(nb_users, nb_movies)
shape = (nb_users, nb_movies)
return shape
def __init__(self, dataset, sets_parameters, stability_parameters):
super().__init__(dataset)
self.nb_users, self.nb_movies = global_parameters(
sets_parameters=sets_parameters)[0:2]
self.differences = stability_parameters['differences']
self.probability = stability_parameters['probability']
self.rmse = stability_parameters['rmse']
print('check self.rmse different from zero:')
print(self.rmse)
self.subsets_number = stability_parameters['subsets_number']
self.landa_array = stability_parameters['landa_array']
assert np.size(self.landa_array) == self.subsets_number + 1
self.coefficients = self.run()
self.category_indices = {'user': 0, 'rmse': 1, 'coefficients': 2}
self.category_matrix = {
'user': self.ratings,
'rmse': self.ratings,
'coefficients': self.coefficients
}
self.category_permute = {
'user': self.permute,
'rmse': self.identity,
'coefficients': self.permute
}
def __init__(self, dataset, sets_parameters, stability_parameters):
super().__init__(dataset)
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
self.differences = stability_parameters['differences']
self.probability = stability_parameters['probability']
self.rmse = stability_parameters['rmse']
print('check self.rmse different from zero:')
print(self.rmse)
self.subsets_number = stability_parameters['subsets_number']
self.landa_array = stability_parameters['landa_array']
print(self.landa_array)
assert np.size(self.landa_array) == self.subsets_number + 1
self.coefficients = self.run()
self.category_indices = {'user': 0, 'rmse': 1, 'coefficients': 2}
self.category_matrix = {'user': self.ratings,
'rmse': self.ratings,
'coefficients': self.coefficients}
self.category_permute = {'user': self.permute,
'rmse': self.identity,
'coefficients': self.permute}
def __init__(self, sets_parameters, Train_set, batch_size, learning_rate0, learning_decay):
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
self.Train_set = Train_set
self.batch_size = batch_size
self.learning_rate0 = learning_rate0
self.learning_rate = learning_rate0
self.learning_decay = learning_decay
def __init__(self, parameters, sets):
self.database = parameters['sets']['database_id']
self.hidden1_units = parameters['autoencoder']['hidden1_units']
self.regularisation = parameters['autoencoder']['regularisation']
self.learning_rate0 = parameters['autoencoder']['learning_rate0']
self.learning_decay = parameters['autoencoder']['learning_decay']
self.batch_size_evaluate = parameters['autoencoder']['batch_size_evaluate']
self.batch_size_train = parameters['autoencoder']['batch_size_train']
self.is_test = parameters['autoencoder']['is_test']
self.nb_users, self.nb_movies = global_parameters(sets_parameters=parameters['sets'])[0:2]
self.difference_matrix = 0
self.rmse = 0
self.rmse_train = 0
self.epoch_steps = int(self.nb_users / self.batch_size_train)
self.nb_steps = parameters['autoencoder']['nb_epoch'] * self.epoch_steps
self.Train_set = Dataset(sets['autoencoder'][0])
self.Validation_set = Dataset(sets['autoencoder'][1])
self.Test_set = Dataset(sets['autoencoder'][2])
self.Loss = Loss()
self.Train = Train(sets_parameters=parameters['sets'],
Train_set=self.Train_set,
batch_size=self.batch_size_train,
learning_decay=self.learning_decay,
learning_rate0=self.learning_rate0)
self.Evaluation = Evaluation(sets_parameters=parameters['sets'],
batch_size_evaluate=self.batch_size_evaluate,
Train_set=self.Train_set)
def __init__(self, sets_parameters):
self.sets_parameters = sets_parameters
self.validation_ratio = sets_parameters['validation_ratio']
self.test_ratio = sets_parameters['test_ratio']
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
self.database_id = sets_parameters['database_id']
self.learning_type = sets_parameters['learning_type']
self.train_val, self.test = self.first_split()
def __init__(self, factorisation_sets, factorisation_parameters, sets_parameters):
if sets_parameters['learning_type'] == 'U':
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
else:
self.nb_movies, self.nb_users = global_parameters(sets_parameters=sets_parameters)[0:2]
self.dimension = factorisation_parameters['dimension']
self.iterations = factorisation_parameters['iterations']
self.landa = factorisation_parameters['landa']
self.train_set = factorisation_sets[0]
self.TrainSet_movies = factorisation_sets[0].transpose(copy=True).tocsr()
self.validation_set = factorisation_sets[1]
self.R = np.empty((self.nb_users, self.nb_movies))
self.U = np.random.rand(self.dimension, self.nb_users)
self.V = np.random.rand(self.nb_movies, self.dimension)
self.rmse, self.difference_matrix = self.run()
def __init__(self, factorisation_sets, factorisation_parameters,
sets_parameters):
if sets_parameters['learning_type'] == 'U':
self.nb_users, self.nb_movies = global_parameters(
sets_parameters=sets_parameters)[0:2]
else:
self.nb_movies, self.nb_users = global_parameters(
sets_parameters=sets_parameters)[0:2]
self.dimension = factorisation_parameters['dimension']
self.iterations = factorisation_parameters['iterations']
self.landa = factorisation_parameters['landa']
self.train_set = factorisation_sets[0]
self.TrainSet_movies = factorisation_sets[0].transpose(
copy=True).tocsr()
self.validation_set = factorisation_sets[1]
self.R = np.empty((self.nb_users, self.nb_movies))
self.U = np.random.rand(self.dimension, self.nb_users)
self.V = np.random.rand(self.nb_movies, self.dimension)
self.rmse, self.difference_matrix = self.run()
def __init__(self, parameters, sets):
self.database = parameters['sets']['database_id']
self.hidden1_units = parameters['autoencoder']['hidden1_units']
self.regularisation = parameters['autoencoder']['regularisation']
self.learning_rate0 = parameters['autoencoder']['learning_rate0']
self.learning_decay = parameters['autoencoder']['learning_decay']
self.batch_size_evaluate = parameters['autoencoder'][
'batch_size_evaluate']
self.batch_size_train = parameters['autoencoder']['batch_size_train']
self.is_test = parameters['autoencoder']['is_test']
self.nb_users, self.nb_movies = global_parameters(
sets_parameters=parameters['sets'])[0:2]
self.difference_matrix = 0
self.rmse = 0
self.rmse_train = 0
self.epoch_steps = int(self.nb_users / self.batch_size_train)
self.nb_steps = parameters['autoencoder']['nb_epoch'] * self.epoch_steps
self.Train_set = Dataset(sets['autoencoder'][0])
self.Validation_set = Dataset(sets['autoencoder'][1])
self.Test_set = Dataset(sets['autoencoder'][2])
self.Loss = Loss()
self.Train = Train(sets_parameters=parameters['sets'],
Train_set=self.Train_set,
batch_size=self.batch_size_train,
learning_decay=self.learning_decay,
learning_rate0=self.learning_rate0)
self.Evaluation = Evaluation(
sets_parameters=parameters['sets'],
batch_size_evaluate=self.batch_size_evaluate,
Train_set=self.Train_set)
def __init__(self, batch_size_evaluate, sets_parameters, Train_set):
self.batch_size_evaluate = batch_size_evaluate
self.Train_set = Train_set
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
def __init__(self, batch_size_evaluate, sets_parameters, Train_set):
self.batch_size_evaluate = batch_size_evaluate
self.Train_set = Train_set
self.nb_users, self.nb_movies = global_parameters(sets_parameters=sets_parameters)[0:2]
def full_import(sets_parameters):
data_file = global_parameters(sets_parameters)[3]
database = np.genfromtxt(data_file, delimiter=',')[:, 0:3]
database[:, 0:2] -= 1
return database