Build confusion matrix for cross validated results? [closed]

Question

I am using python , and I want to know how to build a confusion matrix after I have cross validated my dataset.

If build a confusion matrix at each fold then I have too many confusion matrices. I want one final CM with the right number of cases(not additive) as a final output.

I am currently using cross_val_score from scikit-learn to do my cross validation.

Answer 1

You could compute all the confusion matrices, and then compute the mean and standard deviation for each entry. You could then report a summarized confusion matrix of means $\pm$ standard deviations.

As long as you explain how you have computed the matrix you report, so that your readers can understand what it means, it should be ok!