get_precision_recall

daart.eval.get_precision_recall(true_classes: ndarray, pred_classes: ndarray, background: int | None = 0, n_classes: int | None = None) → dict[source]

Compute precision and recall for classifier.

Parameters:

  • true_classes (array-like) – entries should be in [0, K-1] where K is the number of classes

  • pred_classes (array-like) – entries should be in [0, K-1] where K is the number of classes

  • background (int or NoneType) – defines the background class that identifies points with no supervised label; these time points are omitted from the precision and recall calculations; if NoneType, no background class is utilized

  • n_classes (int, optional) – total number of non-background classes; if NoneType, will be inferred from true classes

Returns:

‘precision’ (array-like): precision for each class (including background class) ‘recall’ (array-like): recall for each class (including background class)

Return type:

dict