奇异线性模型的数据诊断

DATA DIAGNOSIS OF LINEAR MODEL WITH SINGULARITY

论述了奇异线性模型的平差，并证明了ＶＴＱ－Ｖ＝Ｖ１ＴＱ１１－１Ｖ１＝ｍｉｎ．在此基础上，介绍了高杠杆点、异常点和强影响点的诊断．

This paper deals with the adjustment of linear model with singularity, and shows that VTQ-V=V1TQ11-1V1=min. On the basis, diagnosis of high leverage case and outlier and influence case is presented.