Fuzzy inference system with probability factor and its application in data mining

Abstract

In the fuzzy inference system, the construction of the fuzzy rule-base is a key issue. In this paper we provide an identification method for fuzzy model by interpreting the importance factor of each fuzzy rule as the conditional probability of the consequent given the premise. One method of computing the conditional probability is presented. We call this fuzzy model as the fuzzy inference system with probability factor (FISP). One learning process of FISP is also presented in this paper. The application of FISP in time series predication manifests that FISP is very effective. © Springer-Verlag Berlin Heidelberg 2005.