suave_runner.analysis.wilcoxon_analysis

Compare SUAVE experiment metrics using the paired Wilcoxon signed-rank test.

Functions

analyze(systems, correction[, log])	Run every ordered-pair, per-metric comparison of systems.
apply_correction(results, correction)	Set correction and p_adjusted on every result in place.
format_result(result)	Return a one-line, human-readable summary of one comparison.
holm_correction(p_values)	Return Holm-Bonferroni adjusted p-values, in the input's own order.
load_system(managing_system, csv_path)	Load one managing system's accepted rows from a raw campaign CSV.
main([args])	Run the Wilcoxon analysis node.
paired_test(row_system, col_system, row_df, ...)	Run one paired Wilcoxon comparison for one metric between two systems.
rank_biserial(diffs)	Matched-pairs rank-biserial effect size from paired differences.
save_results(results, systems, result_path, ...)	Write the tidy per-comparison CSV and compatibility matrices.

Classes

WilcoxonAnalysis(**kwargs)

ROS front end for the pairwise, run-matched Wilcoxon analysis.