csep.core.poisson_evaluations.paired_t_test

csep.core.poisson_evaluations.paired_t_test(forecast, benchmark_forecast, observed_catalog, alpha=0.05, scale=False)

Computes the t-test for gridded earthquake forecasts.

This score is positively oriented, meaning that positive values of the information gain indicate that the forecast is performing better than the benchmark forecast.

Parameters:

• forecast (csep.core.forecasts.GriddedForecast) – nd-array storing gridded rates, axis=-1 should be the magnitude column

• benchmark_forecast (csep.core.forecasts.GriddedForecast) – nd-array storing gridded rates, axis=-1 should be the magnitude column

• observed_catalog (csep.core.catalogs.AbstractBaseCatalog) – number of observed earthquakes, should be whole number and >= zero.

• alpha (float) – tolerance level for the type-i error rate of the statistical test

• scale (bool) – if true, scale forecasted rates down to a single day

Returns:

csep.core.evaluations.EvaluationResult

Return type:

evaluation_result