HIFS and Probabilistic Similarity Measure-based Intuitionistic Fuzzy Time Series Forecasting Method

Abstract Hesitant intuitionistic fuzzy sets (HIFSs) are more useful than hesitant fuzzy sets (HFSs) and intuitionistic fuzzy sets (IFSs) to model the uncertainty with non-determinism caused by the availability of multiple membership and non-membership grades. In this research paper, we propose the HIFS-based intuitionistic fuzzy time series (IFTS) forecasting method. The proposed method uses adaptive radius clustering for optimal partitioning and incorporates uncertainty with non-determinism in the process of fuzzification of time series data through HIFS. Intuitionistic fuzzy logical relations (IFLRs) used in proposed forecasting method are based on IFSs obtained by aggregating elements of HIFS. Probabilistic λ-cutting algorithm that groups IFSs according to their similarity measure is used in proposed forecasting method to establish IFLRs. Proposed IFTS forecasting method uses a simple computational rule to forecast the outputs and make the process of forecasting simple. Practicability and utility of the suggested forecasting method is shown by implementing it on time series data of the share price of State Bank of India (SBI) at the Bombay Stock Exchange (BSE), India, and the Taiwan Stock Exchange (TAIEX) data. Reduced error measures and valid statistical measure confirm superiority of the proposed method in forecasting of financial time series data.