A Review on Pattern Classification Using Multilevel and Other Fuzzy Min Max Neural Network Classifier

This paper describes a review on the pattern classification using Multilevel Fuzzy Min Max Neural Network classifier (MLF) and different fuzzy min max domain methods. Multilevel fuzzy min max neural network is a supervised learning method and it uses the concept of fuzzy min max method in a multilevel structure to classify the patterns efficiently. To classify the samples in the overlapping regions MLF uses different classifiers with smaller hyperboxes in different levels. MLF selects the best output among all classifiers as a final output of the network. MLF has the capability of learning overlapped region with a single pass through the data. MLF has the highest performance as well as the lowest sensitivity to expansion parameter in comparison of other fuzzy min max domain methods.