Hesitant Intuitionistic Fuzzy Cognitive Map Based Fuzzy Time Series Forecasting Method

Intuitionistic fuzzy cognitive maps (IFCMs) have been proven effective in modeling and forecasting stationary time series with hesitancy and uncertainty. However, challenges persist in dealing when in time varying non stationary time series influenced by dynamic statistical features and availability of multiple intuitionistic fuzzy numbers (IFNs) for node in cognitive map. Hesitant intuitionistic fuzzy set (HIFS) provides an efficient tool that allows associating multiple IFNs time series data. In this study, we present the notion of hesitant intuitionistic fuzzy cognitive map (HIFCM) and propose a HIFCM based fuzzy time series forecasting approach. Weights associated with membership and non-membership grades of IFN are optimized by using particle swarm optimization. With a view to reveal the applicability and utility of the suggested forecasting technique, it is applied to time series data of enrolments of the University of Alabama. The model's performance is gauged in accordance to Root mean square error (RMSE), and a low value shows the model's superiority over other previously existing methods.