On the Equivalence Between AR Family Time Series Models and Fuzzy Models in Signal Processing

摘要

In the paper an advanced analysis of the relationships between statistical Autoregressive (AR) type models and fuzzy models have been presented. The examined family of AR type models includes Autoregressive models of order p, AR(p), Threshold AR (TAR) as well as Smooth Transition Autoregressive (STAR) models. On the other hand, fuzzy models representing different approach, characteristic for Computational Intelligence technics, have been tested for time series analysis and forecasting. The data have been taken from financial market. The research can enrich knowledge which is useful for experts using both approaches to modelling.