Industrial applications of short-term prediction on chaotic time series by local fuzzy reconstruction method

Abstract

The paper describes nonlinear short-term prediction as a possible application of chaos engineering. The authors developed the local fuzzy reconstruction method which is categorized as a nonlinear reconstruction method for nonlinear short-term prediction, and compared prediction performance with linear reconstruction methods, i.e. the Gram-Schumidt orthogonal system method and the tessellation method. The result is that the local fuzzy reconstruction method has advantages in prediction performance and computation time. The authors applied the local fuzzy reconstruction method to practical time series data. The paper considers the local reconstruction method as nonlinear short-term prediction and applications in industrial fields.