The ease of use of fuzzy time series methods and their success in forecasting performance have led to a rapid increase in research in this field. While classical fuzzy time series methods operate solely on membership values, intuitionistic fuzzy time series methods are based on both membership and non-membership values. In this study, a new mixed-order single-variable intuitionistic fuzzy time series forecasting method is proposed. The proposed approach integrates an artificial neural network, the intuitionistic fuzzy c-means algorithm, and the grey wolf optimization algorithm. The intuitionistic fuzzy time series is constructed using crisp values, membership degrees, and non-membership degrees. Fuzzy relationships are determined through a novel artificial neural network based on the dendritic neuron model and optimized using the grey wolf optimization algorithm. Forecasting models are developed separately based on membership and non-membership values, and the final forecasts are obtained by combining these models using weights determined by the grey wolf optimization algorithm. The performance of the proposed method is evaluated and compared with several existing fuzzy time series methods from the literature using different real-world time series datasets.
Cansu et al. (Thu,) studied this question.