摘要
针对输入和输出的模糊数据,根据最小一乘准则,提出了基于对称三角模糊数的模糊最小一乘线性回归模型,并将它转化成线性规划问题,应用Matlab进行求解,通过实例,并与Tanaka,Diamond,Kim-Bishu,Kao-Chyu,Nasrabadi,Chang等模型相比较,验证了我们的模型具有更优的拟合性和良好的稳定性.
Aiming at crisp input and fuzzy output, a certain distance between fuzzy numbers was introduced. Fuzzy least absolute linear regression model based on this distance was introduced and was transformed into linear programming; some theoretical conclusions and algorithm was investigated in detail. Finally, three numerical examples were used to illustrate that the proposed model was reasonable in comparison with Tanaka, Diamond, Kim-Bishu, Kao-Chyu, Nasrabadi and Chang models.
出处
《北京师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第4期348-353,共6页
Journal of Beijing Normal University(Natural Science)
基金
国家自然科学基金资助项目(60775032)
教育部博士点基金资助项目(20020027013)
教育部科学技术重点项目(03184)
国家"九七三"计划资助项目(2002CB312200)
关键词
模糊数据
模糊最小一乘
模糊线性回归
参数估计
fuzzy data
fuzzy least absolute
fuzzy linear regression
parameter estimation
