In the diagonal the number of correctly classified samples are equal to the total number of samples in that particular class but the sensitivity doesn't appear to be 1.000
One thought is that the results in the table or confusion matrix are for calibration whereas the sensitivity and specificity values are for internal cross-validation (CV). What are your sensitivity and specificity values? Are they off by a substantial amount?
Actually the formula for the sensitivity is not satisfied by the result. For classification purpose the formula says (total number of correctly classified samples in a class/total number of samples in that class). If all the samples in a class are correctly classified then the value of sensitivity should be 1.000 but it appears around 0.800. Can you guide me please?