What is the easiest way to obtain the confusion matrix from python CRFSuite?

Question

I am trying to obtain the confusion matrix from python CRFsuite.

This my code:

from sklearn.metrics import confusion_matrix
confusion_matrix(y_test, pred_y, normalize='true', labels=lables)

error:

ValueError: You appear to be using a legacy multi-label data representation. Sequence of sequences are no longer supported; use a binary array or sparse matrix instead - the MultiLabelBinarizer transformer can convert to this format.

I tried to use MultiLabelBinarizer(), but still couldn't get the confusion matrix.

After googling around I found this answer, it says that for the confusion matrix function you have to flatten the y_test and pred_y. I took a look at the source code of CRFsuite for other metrics here, they do use a fallaten function:

def _flattens_y(func):
    @wraps(func)
    def wrapper(y_true, y_pred, *args, **kwargs):
        y_true_flat = flatten(y_true)
        y_pred_flat = flatten(y_pred)
        return func(y_true_flat, y_pred_flat, *args, **kwargs)
    return wrapper

But there is no function for obtaining the confusion matrix.

The y_test and pred_y are nested lists.

How can I flatten the y_test and pred_y to obtain the confusion matrix?

Thank you.

Answer 1

from itertools import chain

f_y_test = list(chain.from_iterable(y_test))
f_pred_y = list(chain.from_iterable(pred_y))