Multi-factors high-order fuzzy time series based on intervals ratio for forecasting

In recent years, fuzzy time series for forecasting have been considerably developed. In this paper, we combine fuzzy time series method for predicting between Lee et al’s method and Huarng and Yu’s method to improve prediction accuracy. Firstly, determine the universe of discourse and the range of intervals using intervals ratio algorithm. Secondly, partition the interval. Thirdly, construct fuzzy logical relationship (FLR) and fuzzy logical relationship group (FLRG). Then, calculate the defuzzyfication. The data for the simulation uses temperature daily data for October 2021 in Semarang, Indonesia. We used daily average temperature data and daily average humidity data as the main factor and a influence factor, respectively for the simulation. The results and error of forecasting values of the proposed method are compared with the previous method. The combination between Lee et al’s method and Huarng and Yu’s method provides better forecasting than the previously using average forecasting error rate (AFER) this show that the proposed method has smaller error than the previous method. The combination method give error value around 0,3%.