Testing Methods

class batcore.tester.TesterBase

Base interface for the tester class

Parameters:

• recommender – recommender to be tested

• data_iterator – iterator over data on which the recommender will be tested.

• kwargs (args,) – any additional params

class batcore.tester.RecTester

tester for the standard recommendations metrics

class batcore.tester.RecTesterAliasTest

tester for the standard recommendations metrics

test_recommender(recommender, data_iterator, flag='self_review', top_ns=None, verbose=False, log_file_path=None, log_stdout=False, *args, **kwargs)

Parameters:

• recommender – recommender to be tested. Must implement RecommenderBase interface

• data_iterator – iterator over dataset on which recommender will be tested. Must implement IteratorBase interface

• top_ns – array of k-s to calculate accuracy@k

Returns:

dictionary with acc@k and mean reciprocal rank, recommendations over time

class batcore.tester.SimulTester

Class for testing non-recommendation metrics on a simulated history

count_metric_dif(metric)

calculates metric on actual and simulated history and returns its relative difference

simulate(recommender, dataset)

Parameters:

• recommender – recommender used to simulate reviewer history

• dataset – dataset for which history will be simulated

Returns:

None. all results are gathered in real and simulated fields