classification (forecasting confusion matrix) ============================================= Forecasting-regime confusion-matrix primitives (v0.2.0). These turn the alarm-vs-seizure match counts into the standard binary-classifier scores (specificity, PPV, NPV, F1) and the observed lead time, under the SOP-length-opportunity true-negative convention documented in :doc:`math/alarm_confusion_matrix` and ADR-0001. This is an internal module (``_classification``); its results are surfaced to users through :doc:`forecasting` (``evaluate`` / ``evaluate_stream``) and the :doc:`report` fields ``specificity`` / ``ppv`` / ``npv`` / ``forecasting_f1`` / ``n_tn`` / ``n_opportunities`` / ``lead_time_mean`` / ``lead_time_median``. It is documented here so the exact TN convention, the NaN-on-empty (fail-loud) behaviour, and the lead-time definition are discoverable in the API reference. .. automodule:: scitex_seizure_metrics._classification :members: :undoc-members: :show-inheritance: