A fuzzy derivative model approach to time-series prediction

TL;DR

This paper introduces a fuzzy system-based method for nonlinear time-series prediction that incorporates derivative information and uses a Taylor ODE solver, demonstrating competitive results on chaotic data.

Contribution

It proposes a novel fuzzy derivative model combined with a Taylor ODE solver for improved nonlinear time-series prediction.

Findings

Achieved comparable or better performance than existing fuzzy and neural network predictors.

Successfully applied to the Mackey-Glass chaotic time-series benchmark.

Demonstrated the effectiveness of derivative-based fuzzy modeling in time-series prediction.

Abstract

This paper presents a fuzzy system approach to the prediction of nonlinear time-series and dynamical systems. To do this, the underlying mechanism governing a time-series is perceived by a modified structure of a fuzzy system in order to capture the time-series behaviour, as well as the information about its successive time derivatives. The prediction task is carried out by a fuzzy predictor based on the extracted rules and on a Taylor ODE solver. The approach has been applied to a benchmark problem: the Mackey- Glass chaotic time-series. Furthermore, comparative studies with other fuzzy and neural network predictors were made and these suggest equal or even better performance of the herein presented approach.