limen.metrics

Score classifier and regressor outputs so experiments can compare permutations on consistent metrics.

Canonical docs

• Standard Metrics Library

What this package owns

Owns the metric helpers used inside model functions and experiment outputs. Does not own model fitting, prediction generation, experiment logging, or backtesting.

Key entry points

Entry point	Use it when	Notes
binary_metrics	You need binary-classification metrics from predictions and probabilities	Import from limen.metrics.binary_metrics for the function form
multiclass_metrics	You need macro or weighted metrics for multiclass problems	Import from limen.metrics.multiclass_metrics
continuous_metrics	You need regression metrics like MAE, RMSE, and R2	Import from limen.metrics.continuous_metrics
safe_ovr_auc	You need OvR AUC without blowing up on missing-class edge cases	Import from limen.metrics.safe_ovr_auc
balanced_metric	You want a single optimization target for balanced binary prediction quality	Exported directly from the package root

Adjacent modules

• limen.sfd.reference_architecture is a common caller of these helpers.
• limen.log and limen.backtest evaluate experiment outcomes after the model phase, but they are downstream from these raw metrics.
• limen.utils re-exports a small subset of metrics for convenience on older code paths.

Quick orientation

metrics/
├── binary_metrics.py
├── multiclass_metrics.py
├── continuous_metrics.py
├── safe_ovr_auc.py
└── balanced_metric.py

Things to know

• The clearest import style is function-level imports from each module, even though the package root also exposes part of the surface.
• binary_metrics assumes the standard Limen data_dict shape and reads data['y_test'].
• safe_ovr_auc returns NaN rather than raising when the class structure makes AUC undefined.
• balanced_metric is Limen-specific and should not be described as a standard sklearn metric.

Read next

• Standard Metrics Library
• Single-File Decoder
• Log