skill_score

hydrostats.ens_metrics.skill_score(scores: float | ndarray[tuple[Any, ...], dtype[floating | integer]], bench_scores: float | ndarray[tuple[Any, ...], dtype[floating | integer]], perf_score: float, eff_sample_size: float | None = None, remove_nan_inf: bool = False) → SkillScoreReturn

Calculate the skill score of the given function.

Parameters:

  • scores – The verification scores, or the mean of the verification scores in an ndarray (float).

  • bench_scores – The reference or benchmark verification scores, or the mean of the benchmark scores (float).

  • perf_score – The perfect score of the score, typically 1 or 0.

  • eff_sample_size – The effective sample size of the data to be used when estimating the sampling uncertainty. Default is None, which will set the eff_sample_size to the length of scores.

  • remove_nan_inf – If True, removes NaN and Inf values in the scores if they exist, pairwise. If False (default), the function will raise an exception.

Returns:

Dictionary containing: {“skillScore”: Float, the skill score, “standardDeviation”: Float, the estimated standard deviation of the skill score} If the scores and bench scores given were floats, the standard deviation will be NaN.

Return type:

dict

References