imultiple_run_metric

Base module defining the interface for multi-run evaluation metrics.

This module provides the generic IMultipleRunMetric source class, establishing the standard evaluation protocol for change point detection algorithms over a series of labeled data providers.

class pysatl_cpd.analysis.metrics.abstracts.imultiple_run_metric.IMultipleRunMetric[source]

Bases: ABC, Generic

Abstract source class for all multi-run evaluation metrics.

Provides a generic interface to evaluate a collection of detection traces against their corresponding labeled ground truth datasets.

abstractmethod evaluate(runs)[source]

Evaluate the detection traces against the provided labeled data.

Parameters:

runs (Sequence[SingleRun[TypeVar(TraceT, bound= DetectionTrace), TypeVar(ProviderT, bound= LabeledData[Any, Any])]]) – A sequence of runs containing a detection trace and the corresponding labeled data provider.

Returns:

The computed metric result over the entire dataset.

Return type:

TypeVar(ResultT)