A novel method for constructing fuzzy classifiers by using SVMs

In this paper, a novel approach to construct fuzzy classifiers by using support vector machines (SVMs) without bias term is proposed. The connection between fuzzy classifiers and support vector classifiers is investigated, and the link between fuzzy rules and kernels is established. It is showed that the proposed method has the inherent advantage that the new fuzzy classifiers do not have to determine the number of rules in advance. Furthermore, the functional equivalence of the two quite different classifiers is proved. The performance of the proposed approach is illustrated by IRIS data sets and comparisons with other methods are also provided.