In multi-label classification, how do we find the P values and confidence interval for the performance evaluation metrics?

Question

I am doing a multilabel classification using two different classifiers with a dataset of 7 labels and 20 features. I have computed the accuracy, sensitivity, specificity, an