fbeta#

openstef_beam.metrics.fbeta(precision_recall: PrecisionRecall, beta: float = 2.0) → float[source]#

Calculate the F-beta score from precision and recall metrics.

The F-beta score is a weighted harmonic mean of precision and recall, allowing for different emphasis on recall versus precision based on the beta parameter.

Parameters:

  • precision_recall (PrecisionRecall) – Container with precision and recall values from the precision_recall function.

  • beta (float) – Weight parameter controlling the relative importance of recall. Values > 1 favor recall, values < 1 favor precision. Common values: 0.5 (favor precision), 1.0 (balanced F1), 2.0 (favor recall).

Returns:

The F-beta score as a float in range [0, 1]. Returns 0.0 if both precision and recall are zero.

Return type:

float

Example

Calculate F2 score (favoring recall):

>>> pr = PrecisionRecall(precision=0.8, recall=0.9)
>>> score = fbeta(pr, beta=2.0)
>>> round(score, 3)
0.878

Calculate F1 score (balanced):

>>> score = fbeta(pr, beta=1.0)
>>> round(score, 3)
0.847

Note

F-beta scores are particularly useful for imbalanced datasets where the cost of false positives and false negatives differs significantly. In energy forecasting, beta > 1 is often preferred to minimize missed peaks.

Parameters:
Return type:

float