节点文献
线性模型中基于稳健诊断的局部影响分析(英文)
Assessing local influenceusing robust diagnostic in linear regression
【摘要】 在线性回归模型的局部影响分析中,通过用具有高崩溃点的稳健估计替代局部影响分析诊断统计量中投影矩阵的方法,给出了一种有效识别影响点的方法.实例分析说明,该方法对高杠杆点及掩盖效应具有较好的识别效果.
【Abstract】 A new method is suggested to study local influence in linear regression model by inserting a high-breakdown point estimation in projection matrix.It is shown that this method is effective to identify the joint influence and masking effects appeared in leverage points.Three examples are analyzed for illustration.
【关键词】 局部影响;
扰动模式;
掩盖效应;
广义Cook统计量;
杠杆点;
稳健诊断;
【Key words】 local influence; perturbation scheme; unmasking; generalized Cook statistic; leverage; robust diagnostics;
【Key words】 local influence; perturbation scheme; unmasking; generalized Cook statistic; leverage; robust diagnostics;
【基金】 ThisresearchissupportedbyNationalNaturalScienceFoundationofChina(10261009)andScienceFoundationofYunnanProvince ( 2 0 0 0A 0 0 0 3M )
- 【文献出处】 云南大学学报(自然科学版) ,Journal of Yunnan University (Natural Sciences) , 编辑部邮箱 ,2004年01期
- 【分类号】O29
- 【被引频次】3
- 【下载频次】59