Struct ABTestAnalyzer 

pub struct ABTestAnalyzer<'a> { /* private fields */ }

Statistical significance testing for A/B experiment results.

Implementations

impl<'a> ABTestAnalyzer<'a>

pub fn new(config: &'a ABTestConfig) -> Self

Create with the test configuration.

pub fn analyze( &self, control_name: &str, control_metrics: &[f64], treatment_name: &str, treatment_metrics: &[f64], ) -> Result<ABTestReport>

Run Welch’s t-test and Mann-Whitney U test, then build a report.

pub fn welchs_ttest(&self, a: &[f64], b: &[f64]) -> Result<TTestResult>

Welch’s t-test (unequal variance, two-sided).

Returns t-statistic, degrees of freedom (Welch-Satterthwaite), and a p-value approximated via Student’s t CDF.

pub fn mann_whitney_u(&self, a: &[f64], b: &[f64]) -> Result<MannWhitneyResult>

Mann-Whitney U test (Wilcoxon rank-sum, two-sided).

Suitable for non-normally distributed metrics. Uses the normal approximation for p-value when n > 20.

pub fn cohens_d(&self, a: &[f64], b: &[f64]) -> f64

Cohen’s d effect size: (mean_a - mean_b) / pooled_std.