confusion_matrix - too many values to unpack

Question

I'm trying to use the confusion_matrix function, as follows:

tn, fp, fn, tp = confusion_matrix(y_true, y_predict).ravel()

y_true and y_predict are both lists. When I return their shape, I get: (71,) .

I'm however getting the following error for the above statement:

ValueError: too many values to unpack

I'm not sure if it is because of the second (empty) dimension in (71,) ? I'm not sure how to remove it if it is the issue here.

Any thoughts?

Thanks.

Answer 1

You can only assign multiple variables dynamically if the number of outputs is certain. If you assign the result of confusion_matrix to a single variable, you can then check its contents in a loop and assign the contents conditionally:

returned = confusion_matrix(y_true, y_predict).ravel()

for var in returned:
    #... do stuff with each item in the returned collection

You could also just check its length and if it is 4, you can proceed as usual:

if len(returned) == 4:
    tn, fp, fn, tp = returned

Answer 2

The line you're trying to execute tn, fp, fn, tp = confusion_matrix(y_true, y_predict).ravel() is valid only if you have 2 classes in output (binary classification). However, the error you get is an indicator that you have more than 2 classes (multi-class classification). In this case, there's no meaning of tn, fp, fn, tp . Instead you can visualize the confusion matrix (eg by using heatmap).