mcframework.stats_engine.percentiles

mcframework.stats_engine.percentiles(x: ndarray, ctx: StatsContext) → dict[int, float]

Empirical percentiles evaluated on the cleaned sample.

For each \(p \in \mathcal{P}\) we compute the empirical quantile \(Q_p(x)\) using numpy.percentile() (linear interpolation).

Parameters:

xndarray

Input sample.

ctxStatsContext

Uses percentiles and nan_policy.

Returns:

dict[int, float]

Mapping \(p \mapsto Q_p(x)\).

Examples

>>> percentiles(np.array([0., 1., 2., 3.]), {"percentiles": (50, 75)})
{50: 1.5, 75: 2.25}