A TSK fuzzy system based on the rate of change of moving average for time series prediction

It is very difficult for nonstationary or nonlinear time series to establish a proper prediction model. In this paper, we propose a TSK FPS that uses the rate of change of MA of data as the predictor input. The LPF property of the MA stabilizes statistical characteristics of time series and the rate of change of them removes the monotonic increasing/decreasing tendency of the data, thereby overcoming the rule bias of the data. For the fuzzy partitioning of input space, CBKM using the correlation coefficient as the similarity index is applied. As a result, more appropriate rules to the attribute of data can be obtained. In order to verify the validity of the proposed method, simulations were performed on representative cases of nonstationary and nonlinear time series, respectively. The result of the simulation shows excellent performance and effectiveness of the proposed method. Copyright © The Korean Institute of Electrical Engineers.