From Epanechnikov Mixture Model to Rule-Centered Fuzzy Model and its Application to Stock Index Prediction
Qin Li Zhang · Advanced materials research · 2014
In this paper, the mathematical equivalence between the conditional mean of a Epanechnikov mixture model (EMM) and the defuzzified output of a rule-centered generalized fuzzy model (RCGFM) is derived theoretically. Our results provide a new perspective for fuzzy systems, i.e., interpreting them from a probabilistic viewpoint. Thus, instead of directly estimating the parameters of the fuzzy rules in a rule-centered generalized fuzzy model, we can first estimate the parameters of the corresponding EMM using any popular density estimation algorithm like the expectation maximization (EM) algorithm. Our experimental results clearly indicate that a rule-centered generalized fuzzy model trained in such a new way has higher approximation accuracy and generalization ability than other models.