A Deterministic Nonlinear Prediction Method through Fuzzy Reasoning Using Neighborhoods' Difference and Its Application to Actual Time Series Data

Abstract

In this paper, we focus on chaos time series analysis studied widely and actively in recent years and propose a new deterministic nonlinear prediction technique of time series through fuzzy reasoning using neighborhoods' difference in the reconstructed state space based on the Takens's theorem. Then, the proposed method is compared with linear autoregressive prediction method and local fuzzy reconstruction method, which is one of deterministic nonlinear time series prediction methods through their application to actual time series data and its efficiency is demonstrated.