This paper exhibits the connection between fuzzy inference system and kernel machines, and proposes a support vector learning approach to construct fuzzy inference system so that it can have good generalization ability in a high dimensional feature space. It is showed that the two seemingly unrelated research areas, fuzzy inference systems and kernel machines, are closely related. Under some minor constrains, the equivalence of the two seemingly quite distinct models is proved. The designed fuzzy inference system can be represented as a decision function consisting of series expansion of modified fuzzy basis functions (MFBFs), and this also makes itself to be interpretable. The approach preserves advantages of both the statistical learning framework and the fuzzy inference system. The performance of the proposed approach is illustrated by an example of nonlinear function regression.
Kernel method for constructing fuzzy inference system
2006-01-01
1366792 byte
Conference paper
Electronic Resource
English
A New Method for Constructing Kernel Function of Support Vector Regression
British Library Online Contents | 2006
|