pyconstruct.learners.StructuredPerceptron.score

StructuredPerceptron.score(X, Y, Y_pred=None, **kwargs)

Compute the score as the average loss over the examples.

This method is needed for scikit-learn estimation in GridSearchCV and other model selection methods.

Parameters:

• X (numpy.ndarray) – An array of input examples. The first dimension must be the number of samples.
• Y (numpy.ndarray) – An array of true output objects.
• Y_pred (numpy.ndarray) – An array of predicted object.

Returns:

score – The score of the model over the examples.

Return type:

float