Design of grey model selector

• Main Authors: Cheng Hau-Chian, 鄭皓謙
• Other Authors: Hsu Yen-Tseng
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/76225407805479960289

碩士 === 國立臺灣科技大學 === 電機工程技術研究所 === 85 === The original data are vary significantly in the long-term prediction. But almost of all prediction model has a characteristic that is only used in a few of distributions in our recognition for prediction. Therefore, a single model is used to predict, its results are not good enough in long-term prediction. This paper propose a novel method called grey model selector to solve this problem and to obtain the results accurately. According to the characteristic of the original data, we have separated it into two kinds of typical sequence in this paper, such as monotone sequence and central symmetry sequence. On the basis of the grey system, this paper uses the result of fuzzy tendency prediction to decide the appropriate grey model which has the less prediction error. The proposed method has been applied to model a non-linear system - Mackey-Glass chaotic time series, and compared with combined model of grey model and Wang-Mendel approach. The simulation results show that this novel method has more accurate than the other approaches