Let us suppose I have a NxN matrix and I want to classify in M classes each entry of the matrix using a fuzzy classifier. The output of my classifier will be, for each matrix entry, an M-dimensional vector containing the probabilities for the entry to be classified in each class. A naive way to build a confusion matrix would be to select the highest probability in each vector and use it as a crips classification. However, I would like to take into account all the probabilities associated to each entry and compute a "fuzzy" confusion matrix. Is this possible?



You must log in to answer this question.

Browse other questions tagged .