Building a hierarchical representation of membership functions

Deriving inference rules from training examples is one of the most common machine-learning approaches. Fuzzy systems that can automatically derive fuzzy if-then rules and membership functions from numeric data have recently been developed. In this paper, we propose a new hierarchical representation for membership functions, and design a procedure to derive them. Experimental results on the Iris data show that our method can achieve high accuracy. The proposed method is thus useful in constructing membership functions and in managing uncertainty and vagueness.