precision_recall#

openstef_beam.metrics.precision_recall(cm: ConfusionMatrix, *, effective: bool = False) → PrecisionRecall[source]#

Calculate precision and recall metrics from a confusion matrix.

These metrics evaluate the quality of peak detection by measuring how many predicted peaks were correct (precision) and how many actual peaks were detected (recall).

Parameters:

  • cm (ConfusionMatrix) – Confusion matrix from the confusion_matrix function containing all classification outcomes.

  • effective (bool) – If True, uses effective true positives which account for prediction direction and magnitude. If False, uses standard true positives for calculation.

Returns:

PrecisionRecall containing precision and recall values, each in range [0, 1].

Return type:

PrecisionRecall

Example

Calculate standard precision and recall:

>>> import numpy as np
>>> y_true = np.array([100, 150, 80, 200, 90])
>>> y_pred = np.array([105, 145, 85, 195, 95])
>>> cm = confusion_matrix(y_true, y_pred, limit_pos=120, limit_neg=85)
>>> pr = precision_recall(cm)
>>> float(pr.precision)
1.0
>>> float(pr.recall)
1.0

Note

When effective=True, the metrics focus on predictions that provide actionable information (correct magnitude and direction) rather than just correct classification.

Parameters:
Return type:

PrecisionRecall