@revista_internacional{1033,
  keywords = {hesitant fuzzy set, Fuzzy time series, Improved cumulative probability distribution approach, Weighted fuzzy logic relationship, High-order model forecasting},
  author = {Chuyi Zhang and Deshan Sun and Kuo Pang and Li Zou and Luis Martínez and Witold Pedrycz},
  title = {A high-order hesitancy fuzzy time series model based on improved cumulative probability distribution approach and weighted fuzzy logic relationship},
  abstract = {Fuzzy time series models, with their unique capability to handle uncertainty, have become crucial tools in managing complex and imprecise data environments. The proposal of hesitant fuzzy set provides an effective solution for addressing the uncertainty encountered when determining membership degrees in time series data. To enhance the credibility and forecasting accuracy of fuzzy time series model, this paper proposes a high-order hesitant fuzzy time series model based on an improved cumulative probability distribution approach (ICPDA) and weighted fuzzy logic relationship. First, the distribution characteristics and dispersion degrees of time series data are more comprehensively considered by refining the cumulative probability distribution approach with statistical measures, achieving a more adaptive partitioning of time series data. Second, triangular and Gaussian membership functions are employed to construct hesitant fuzzy sets, which are then aggregated using aggregation operators to define fuzzy time series, effectively capturing multiple uncertainties inherent in time series data. In addition, to further improve the forecasting capability of the model, weights are integrated into the fuzzy logical relations, facilitating the defuzzification output for forecasts. The model is finally applied to two real-world time series datasets: the enrollment numbers of the University of Alabama and car sells of Quebec City in Canada, with comparative experiments demonstrating the model s strong generalization ability and forecasting accuracy in uncertain environments.},
  year = {2025},
  journal = {Information Sciences},
  volume = {717},
  pages = {122262},
  issn = {0020-0255},
  url = {https://www.sciencedirect.com/science/article/pii/S0020025525003949},
  doi = {https://doi.org/10.1016/j.ins.2025.122262},
}