Changelog

All notable changes to this project will be documented in this file.

[1.0.0] - Upcoming

Added

• Core evaluation engine (EvaluationContext).
• Eight built-in Doctors (OverfittingDoctor, LeakageDoctor, PredictionDoctor, DataDoctor, FeatureDoctor, CalibrationDoctor, ProductionDoctor, GeneralizationDoctor).
• Evidence, Confidence, Risk, and Prescription engines.
• HTML Dashboard rendering.
• MLflow Integration (log_report).
• Validation Laboratory framework featuring 54 benchmark scenarios.

Changed

• Refactored BaseDoctor to use the EvidenceBuilder.
• Tuned global thresholds in rules.py for 98.1% diagnostic accuracy.

Removed

• Deprecated legacy hardcoded boolean checks in favor of the normalized Evidence Engine.