Intuitionistic Fuzzy Neural Network for Time Series Forecasting - The Case of Metal Prices

Abstract

Forecasting time series is an important problem addressed for years. Despite that, it still raises an active interest of researchers. The main issue related to that problem is the inherent uncertainty in data which is hard to be represented in the form of a forecasting model. To solve that issue, a fuzzy model of time series was proposed. Recent developments of that model extend the level of uncertainty involved in data using intuitionistic fuzzy sets. It is, however, worth noting that additional fuzziness exhibits nonlinear behavior. To cope with that issue, we propose a time series model that represents both high uncertainty and non-linearity involved in the data. Specifically, we propose a forecasting model integrating intuitionistic fuzzy sets with neural networks for predicting metal prices. We validate our approach using five financial multivariate time series. The results are compared with those produced by state-of-the-art fuzzy time series models. Thus, we provide solid evidence of high effectiveness of our approach for both one-and five-dayahead forecasting horizons.

Neural intuitionistic fuzzy system with justified granularity

“…• IFNN-TS (intuitionistic fuzzy neural network for TS forecasting) (Hajek et al, 2021), an earlier version of the neural intuitionistic fuzzy system. The parameters of the membership functions and rule antecedents of the IFNN-TS system were set using the subtractive clustering algorithm and the consequent parameters of the if-then rules were adapted using the gradient descent algorithm.…”

“…In addition, intuitionistic fuzzy operators are exploited to obtain the firing weights of if-then rules and a weighted average method intuitionistic fuzzy sets is used to defuzzify the outcome of the proposed neural intuitionistic fuzzy system of Takagi-Sugeno-Kang (TSK) type. In the earlier version of this paper (Hajek et al, 2021), we argued that the main limitation of the used fuzzy clustering approach is that no if-then rules were matched for many observations due to the high volatility in the TS data. To address this issue, here we replace the fuzzy clustering algorithm with fuzzy association rules for rule generation.…”

A Novel Learning Rule in Intuitionistic Fuzzy Artificial Neural Network and its Applications to MAGDM Support System

“…The author in [20] proposed the sensitivity analysis for artificial neural networks and in [22] was proposed various aggregation operations for artificial neural networks. The authors in [7] utilized Intuitionistic fuzzy neural networks for time series forecasting. Authors in [11] proposed varieties of linguistic intuitionistic fuzzy distance measures for Linguistic TOPSIS method for a better decision support system in the field of MAGDM.…”