relaxed.metrics package

relaxed.metrics.asimov_sig(s: Array, b: Array) → float

Median expected significance for a counting experiment, valid in the asymptotic regime. Also valid for the multi-bin case.

Parameters:

• s (Array) – Signal counts.

• b (Array) – Background counts.

Returns:

The expected significance.

Return type:

float

relaxed.metrics.gaussianity(model: PyTree, bestfit_pars: dict[str, ArrayLike], data: Array, rng_key: Any, n_samples: int = 1000) → Array