Author(s): Amir Noori; Hossein Bonakdari

Linked Author(s): Hossein Bonakdari

Keywords: Fuzzy Time Series (FTS); Intuitionistic Fuzzy Set (IFS); Triangular Fuzzy Numbers (TFN); Golden ratio algorithm; Rainfall

Abstract: In this paper, the Fuzzy time series (FTS) model, based on Fuzzy sets, was developed to account for the degree of non-determinacy. In order to solve the hesitancy in the decision-making system, intuitionistic fuzzy sets (IFS) built on FTSs have been organized. In addition, FTSs discourse universe, interval length, and membership functions are determined using induced fuzzy sets and triangular fuzzy numbers (TFN). Furthermore, in the FTS as mentioned above procedures, a weighting technique is used to demonstrate the precise reflection of each specific fuzzy relationship in forecasting and improve forecasting accurateness. Another characteristic of this method is that to improve the accuracy rate of the forecasted values, the predicted values are compared with actual ones using the Golden Ratio Algorithm. Furthermore, for checking the feasibility and predictability of the suggested approach, the improved forecasting model has been used to estimate historical student enrollments at the University of Alabama as a benchmark time series data. In order to represent, the acquired results are contrasted with other existing FTS forecasting models utilizing MSE and AFE measures. The results illustrate that the proposed model has more favorable forecasting accuracy levels than others. The developed forecasting approach is also evaluated and validated using short-time synoptic time series data. Therefore, in a real case study, the proposed model is applied to forecast monthly rainfall for Ottawa City, Ontario (an Eastern Canadian province). According to the empirical findings, the weighted model outperforms one of the conventional FTS models in predicting future time-series data.

DOI: https://doi.org/10.3850/978-90-833476-1-5_iahr40wc-p0235-cd

Year: 2023