latent.stats.intervals¶

Wilson score confidence intervals for binomial proportions.

Reference: Wilson (1927). Delegates to statsmodels.stats.proportion.proportion_confint.

Functions¶

wilson_ci(successes: int, total: int, confidence_level: float = 0.95) -> MetricResult

Wilson score interval for a binomial proportion.

Asymmetric interval that behaves correctly near 0% and 100%.

Reference: Wilson (1927), via statsmodels.stats.proportion.proportion_confint.

Args: successes: Number of successes. total: Total number of trials. confidence_level: Confidence level (default 0.95).

Returns: MetricResult with Wilson interval bounds.