Classification of Imbalanced and Overlapping Classes using Intuitionistic Fuzzy Sets

This paper describes works in progress on the problem of classification of imbalance and overlapping classes using intuitionistic fuzzy sets. A fuzzy set approach is presented first - the classes are recognized using a fuzzy classifier. Next, an intuitionistic fuzzy set representation is proposed. We show that intuitionistic fuzzy classifier has some natural tendency to deal efficiently with imbalanced and overlapping data. We also explore in details the evaluation of the classifier results (especially from the point of view of recognizing the smaller class)