Principle:Interpretml Interpret Classification Performance Curves
Metadata
| Field | Value |
|---|---|
| Sources | Interpretml_Interpret |
| Domains | Machine_Learning, Model_Evaluation |
| Updated | 2026-02-07 |
Overview
Classification performance curves (ROC and Precision-Recall) provide visual evaluation of binary classification model performance across decision thresholds.
Description
ROC (Receiver Operating Characteristic) and PR (Precision-Recall) are performance evaluation explainers that produce threshold-dependent performance curves for binary classification models. ROC curves plot true positive rate against false positive rate, while PR curves plot precision against recall. Both conform to the InterpretML perf explainer API and produce interactive Plotly visualizations with AUC (Area Under Curve) metrics.
Usage
Use ROC curves for balanced datasets and PR curves for imbalanced datasets to evaluate and compare binary classification model performance across all possible decision thresholds.