Function fanova_ranking 

pub fn fanova_ranking(
    view: &StudyView,
    objective: usize,
) -> Result<Vec<(String, f64)>, FitError>

The forest-fANOVA importance ranking of a study’s parameters, most important first.

Reads the completed trials carrying a finite value for objective, builds the column-major fANOVA input over the intersection search space (the parameters present with an identical recorded distribution in every such trial — the same space PED-ANOVA scores over), fits a seeded random forest, and returns each parameter’s quantify_importance(&[i]).mean. Ties break by name, so the order is total and deterministic.

An empty ranking is returned when there is no shared search space (no common parameter across the completed trials), rather than an error.

Errors

[FitError] if the forest cannot be fitted — in practice only when there are no rows to fit (every completed trial lacked a finite objective).