VSE

mlquantify.metrics.VSE(prev_real, prev_pred, train_values)

Compute the Variance-normalised Squared Error (VSE).

Parameters:

prev_realarray-like

True regression values (from test set).

prev_predarray-like

Predicted regression values (from test set).

train_valuesarray-like

True regression values from training set, used to compute variance normalization.

Returns:

verrorfloat

Variance-normalised squared error.