Research on fuzzy modeling forecasting method based on stochastic neural network

Abstract

In this article, a novel modeling forecasting method based on the combination of the Modified Takagi and Sugeno (MTS) fuzzy model and the stochastic neural network is presented. Expectation–Maximization (EM) algorithm is put forward to calculate the parameters of neural network structure and its weights. Theoretical analysis and prediction examples all show that the technique has strong universalized capabilities and the methods are effective.