RNOD

mlquantify.metrics.RNOD(prev_pred, prev_real, distances=None)

Compute the Root Normalised Order-aware Divergence (RNOD) for ordinal quantification evaluation.

Parameters:

prev_realarray-like or dict

True prevalence values for each ordered class.

prev_predarray-like or dict

Predicted prevalence values for each ordered class.

distances2D array-like of shape (n_classes, n_classes), optional

Distance matrix between classes (d(y_i, y_j)). If None, assumes d(y_i, y_j) = |i - j|.

Returns:

rnodfloat

Root Normalised Order-aware Divergence between predicted and true prevalences.