Concepts

Concept pages explain how Model Auditor v0.1.16 behaves across workflows.

• Evaluation model follows data from a pandas DataFrame to subgroup metrics.
• Thresholds and predictions explains scalar, optimized, and conditional operating points.
• Bootstrap confidence intervals describes sampling units, percentile bounds, and reproducibility.
• Results and exports separates numeric result objects, display tables, and styled output.
• Error-group analysis explains the level-versus-rest odds ratio for TP, TN, FP, and FN groups.
• Hierarchical plot data explains branch construction and external rendering.

For exact signatures and defaults, use Reference. For procedures, use Guides.