摘要
目的探讨主成分回归模型中存在异常点时的稳健诊断方法。方法采用基于MVT和LMS方法的一种稳健主成分回归方法来诊断异常点,并结合实例进行方法的对比。结果与未去除异常点时得到的回归方程相比较,具有较小的残差平方和,拟合效果较好。解决了主成分回归中存在异常点的问题。结论当主成分回归中存在异常点时,本文中所述的稳健诊断方法具有较高的稳健性,在主成分回归分析中有较好的应用前景。
Objective To explore the robust diagnostic method when there are outliers in principal component regression. Methods We introduce a robust principal component regression based on MVT and LMS to detect outliers, and compare methods using a practical example. Results The model of the data in which outliers is detected and removed has a lower PRESS and a better fitness than that of the original data It is better to solve the outliers problem in PCR. Conclusion The robust diagnostic method in principal component regression has high robustness when there are outliers in PCR and should be widely used.
出处
《中国卫生统计》
CSCD
北大核心
2008年第1期31-34,共4页
Chinese Journal of Health Statistics
基金
山西省自然基金资助项目(项目编号20021104)
关键词
MVT方法
LMS估计
稳健主成分回归
异常点诊断
MVT method
LMS estimator
robust principal component regression
outlier diagnosis