High-order Type-2 fuzzy time series

Abstract

This paper proposes a novel improved model for fuzzy time series forecasting using high-order and Type-2 fuzzy time series. Some Type-1 fuzzy time series model based on first order and high-order fuzzy time series have already been developed, however the accuracy of the forecasting values sill needs to be improved. In this paper, we present a high-order fuzzy time series model using Type-2 fuzzy sets. The employment of Type-2 fuzzy sets enables the proposed method to achieve a better result compared with existing fuzzy time series models. Our experimental results demonstrate the improved forecasting accuracy and reliability in comparison with other existing fuzzy time series models.