import alpenglow.Getter as rs
import alpenglow as prs
[docs]class BatchAndOnlineFactorExperiment(prs.OnlineExperiment):
"""BatchAndOnlineFactorExperiment(dimension=10,begin_min=-0.01,begin_max=0.01,batch_learning_rate=0.05,batch_regularization_rate=0.0,batch_negative_rate=70,online_learning_rate=0.05,online_regularization_rate=0.0,online_negative_rate=100,period_length=86400)
Combines BatchFactorExperiment and FactorExperiment by updating
the model both in batch and continously.
Parameters
----------
dimension : int
The latent factor dimension of the factormodel.
begin_min : double
The factors are initialized randomly, sampling each element uniformly from the interval (begin_min, begin_max).
begin_max : double
See begin_min.
batch_learning_rate : double
The learning rate used in the batch stochastic gradient descent updates.
batch_regularization_rate : double
The coefficient for the L2 regularization term for batch updates.
batch_negative_rate : int
The number of negative samples generated after each batch update. Useful for implicit recommendation.
timeframe_length : int
The size of historic time interval to iterate over at every batch model retrain. Leave at the default 0 to retrain on everything.
online_learning_rate : double
The learning rate used in the online stochastic gradient descent updates.
online_regularization_rate : double
The coefficient for the L2 regularization term for online update.
online_negative_rate : int
The number of negative samples generated after online each update. Useful for implicit recommendation.
"""
def _config(self, top_k, seed):
model = rs.FactorModel(**self.parameter_defaults(
begin_min=-0.01,
begin_max=0.01,
dimension=10,
initialize_all=False,
))
#
# batch
#
# updater
batch_updater = rs.FactorModelGradientUpdater(**self.parameter_defaults(
learning_rate=self.parameter_default('batch_learning_rate', 0.05),
regularization_rate=0.0
))
batch_updater.set_model(model)
# objective
point_wise = rs.ObjectiveMSE()
batch_gradient_computer = rs.GradientComputerPointWise()
batch_gradient_computer.set_objective(point_wise)
batch_gradient_computer.set_model(model)
batch_gradient_computer.add_gradient_updater(batch_updater)
# negative sample generator
batch_negative_sample_generator = rs.UniformNegativeSampleGenerator(**self.parameter_defaults(
negative_rate=self.parameter_default('batch_negative_rate', 70),
initialize_all=False,
seed=67439852,
filter_repeats=False,
))
batch_negative_sample_generator.add_updater(batch_gradient_computer)
batch_offline_learner = rs.OfflineIteratingOnlineLearnerWrapper(**self.parameter_defaults(
seed=254938879,
number_of_iterations=3,
shuffle=True,
))
batch_offline_learner.add_iterate_updater(batch_negative_sample_generator)
batch_online_learner = rs.PeriodicOfflineLearnerWrapper(**self.parameter_defaults(
write_model=False,
read_model=False,
clear_model=False,
learn=True,
base_out_file_name="",
base_in_file_name="",
))
batch_online_learner.set_model(model)
batch_online_learner.add_offline_learner(batch_offline_learner)
batch_data_generator_parameters = self.parameter_defaults(
timeframe_length=0,
)
if(batch_data_generator_parameters['timeframe_length']==0):
print("Full experiment")
batch_data_generator = rs.CompletePastDataGenerator()
else:
print("Timeframe experiment")
batch_data_generator = rs.TimeframeDataGenerator(**batch_data_generator_parameters)
batch_online_learner.set_data_generator(batch_data_generator)
batch_period_computer = rs.PeriodComputer(**self.parameter_defaults(
period_length=86400,
start_time=-1,
period_mode="time",
))
batch_online_learner.set_period_computer(batch_period_computer)
#
# online
#
# updater
online_updater = rs.FactorModelGradientUpdater(**self.parameter_defaults(
learning_rate=self.parameter_default('online_learning_rate', 0.2),
regularization_rate=0.0
))
online_updater.set_model(model)
# objective
point_wise = rs.ObjectiveMSE()
online_gradient_computer = rs.GradientComputerPointWise()
online_gradient_computer.set_objective(point_wise)
online_gradient_computer.set_model(model)
online_gradient_computer.add_gradient_updater(online_updater)
# negative sample generator
online_negative_sample_generator = rs.UniformNegativeSampleGenerator(**self.parameter_defaults(
negative_rate=self.parameter_default('online_negative_rate', 100),
initialize_all=False,
seed=67439852,
filter_repeats=False,
))
online_negative_sample_generator.add_updater(online_gradient_computer)
learner = [batch_online_learner, online_negative_sample_generator]
return (model, learner, [])