Improved time-variant fuzzy time series forecast |
| |
Authors: | Hao-Tien Liu Nai-Chieh Wei Chiou-Goei Yang |
| |
Affiliation: | (1) Department of Industrial Engineering and Management, I-Shou University, No. 1, Sec. 1, Syuecheng Rd., Dashu Township, Kaohsiung County, 840, Taiwan |
| |
Abstract: | Since Song and Chissom (Fuzzy Set Syst 54:1–9, 1993a) first proposed the structure of fuzzy time series forecast, researchers have devoted themselves to related studies. Among these studies, Hwang et al. (Fuzzy Set Syst 100:217–228, 1998) revised Song and Chissom’s method, and generated better forecasted results. In their method, however, several factors that affect the accuracy of forecast are not taken into consideration, such as levels of window base, length of interval, degrees of membership values, and the existence of outliers. Focusing on these factors, this study proposes an improved fuzzy time series forecasting method. The improved method can provide decision-makers with more precise forecasted values. Two numerical examples are employed to illustrate the proposed method, as well as to compare the forecasting accuracy of the proposed method with that of two fuzzy forecasting methods. The results of the comparison indicate that the proposed method produces more accurate forecasting results. |
| |
Keywords: | Fuzzy time series Time-variant forecast Length of interval Window base Outlier |
本文献已被 SpringerLink 等数据库收录! |
|