ce-explain-interact

What problem does it solve?

Post-process CalibratedExplanations by narrating results, generating plots, inspecting rules, and applying filters to highlight the most relevant features and rules.

Core Features & Use Cases

• Narrated narratives for CalibratedExplanation instances to improve interpretability.
• Plotting and visualization of rule importance, calibrated uncertainty, and feature weights.
• Filtering atomic and conjunctive rules, inspecting both single explanations and collections across factual and alternative explanation types.
• Use Case: data scientists review explanations to quickly identify which features most influence calibrated predictions and communicate findings to stakeholders.

Quick Start

Provide narrated CE explanations with interactive insights for a given CalibratedExplanation.