I have one question to Max Likelihood discriminant values problem after classifikation: wenn i need true probabilities, how can i do that? because in Rule classifier for Rule images of Maximum Likelihood classification file i have only negative values as discriminant values...i have heard, there is possibility to convert this values in true probabilities via Band math with EXP- Function. Can someone tell me if there is any kind of solution? how can i get the new classification file using Rule classifier with probabilities, so that maybe my confusion matrix accuracy will show better results? Thank you very much in advance!