基于离中度的模糊线性回归模型的统计诊断  

Statistical diagnosis of fuzzy linear regression model based on fuzzy decentring degree

作　　者：徐嘉薇 许若宁[1] 

机构地区：[1]广州大学数学与信息科学学院,广东广州510006

出　　处：《广州大学学报（自然科学版）》2014年第6期5-11,共7页Journal of Guangzhou University:Natural Science Edition

基　　金：国家自然科学基金资助项目(11271096)

摘　　要：针对实数值输入、不确定值输出的模糊线性回归模型,讨论了基于模糊离中度数据删除的模型参数估计,建立了观测数据中异常点或强影响点的检验统计诊断量,并给出运用该诊断量进行检验的步骤.文章中的方法不仅能够在一定程度上克服已有方法的某些不足,而且能有效地提高计算结果的准确性,最后实例表明,该统计诊断量是合理有效的.Aimed at the fuzzy linear regression model with real input and fuzzy output,the model parameter estimation is disscussed on data deletion of fuzzy decentring degree.Statistical diagnostic quantity is established to test oudiers or strong influential points in the observed data.And the step is given to examine with the statistical diagnostic quantity.The method can overcome some shortcomings of the existing method in the certain extent,and also can improve the accuracy of the computational results effectively.At last the example suggests that the statistic constructed in this paper is effective.

关 键 词：模糊线性回归模型 数据删除模型 模糊离中度 

分 类 号：O159[理学—数学]