Using Fuzzy Mathematical Programming to Solve Two Step Fuzzy Regression Model

• Main Authors: Sun Ying-Chien, 孫瑛鍵
• Other Authors: Hsu Sung-may
• Format: Others
• Language: zh-TW
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/03663990216721499625

碩士 === 國立臺灣科技大學 === 工業管理系 === 88 === Abstract In 1982, Professor Tanaka produced the concept of fuzzy regression analysis. Because the fuzzy regression model was developed by the possibility theory, it is good to estimate and describe variables when data themselves are not complete or imprecise. During the process of model solving , however, because the reason of constraints, the fuzzy regression model was affected by outliers easily. In order to remove this defect, under the basis structure of fuzzy regression analysis, in addition to my viewpoints, I also refer to some scholarly theses to make a new mathematical calculation. In this thesis, I describe and explain the fuzzy least square regression and point out its defects, and add fuzzy linear programming to reduce the effect of outliers to model. Key words：Fuzzy theory、Regression analysis、Fuzzy least square、 Fuzzy linear programming.