On Generalized Fuzzy Jensen-Exponential Divergence and Its Application to Pattern Recognition

摘要

This paper develops a novel information theoretic divergence measure between two fuzzy sets based on exponential function and applies it to solve pattern recognition problems. First, we generalize the idea of fuzzy Jensen-exponential divergence and propose a new parametric divergence called fuzzy Jensen-exponential divergence of order-α to measure the information of discrimination between two fuzzy sets. We also prove some properties of the proposed measure and discuss its particular cases. Finally, we apply the proposed divergence measure between fuzzy sets to deal with pattern recognition problems with fuzzy information.