Neural intuitionistic fuzzy system with justified granularity

Abstract

Fuzzy systems are commonly used to model uncertainties occurring in data. They have been intensively investigated and extended to be used for the construction of time series forecasting models. In particular, intuitionistic fuzzy sets have been used to capture higher levels of uncertainty occurring in the modeled data. Neural networks were also used to reflect non-linearity relationships frequently observed in time series. Following this line of research, in this paper, we propose a new hybrid fuzzy system merging the most promising data representation methods with an advanced optimization technique, which is the principle of justified granularity. In this technique, we construct an innovative time series forecasting model. In the experimental part of the Article Title paper, we demonstrate the advantages arising from applying the proposed approach to metal price forecasting. We provide solid evidence that the proposed model is competitive with other state-of-the-art models known from the literature for forecasting horizons of one and five days.