Development of Membership Functions
Timothy J. Ross · 2010
This chapter describes a few procedures to develop these membership functions based on deductive intuition or numerical data. Since the membership function essentially embodies all fuzziness for a particular fuzzy set, its description is the essence of a fuzzy property or operation. The chapter also describes six procedures that have been used to build membership functions. The following is a list of six straightforward methods described in the literature to assign membership values or functions to fuzzy variables. The six methods are: intuition, inference, rank ordering, neural networks, genetic algorithms, and inductive reasoning. The chapter illustrates each of these methods illustrated in simple examples. Intuition involves contextual and semantic knowledge about an issue; it can also involve linguistic truth values about this knowledge. In the inference method the author uses knowledge to perform deductive reasoning. The chapter explains how a neural network can be used to determine membership functions. Controlled Vocabulary Terms ART neural nets; genetic algorithms; inference mechanisms