On the mean squared error of the ridge estimator of the covariance and precision matrix

详细信息    查看全文

• 作者：Wessel N. van Wieringen ; ^{wvanwie@few.vu.nl} ; ^{w.vanwieringen@vumc.nl}
• 关键词：Inverse covariance matrix ; Multivariate normal ; ℓ_{2ℓ2-penalization}
• 刊名：Statistics & Probability Letters
• 出版年：2017
• 出版时间：April 2017
• 年：2017
• 卷：123
• 期：Complete
• 页码：88-92
• 全文大小：428 K
• 卷排序：123

文摘

For a suitably chosen ridge penalty parameter, the ridge regression estimator uniformly dominates the maximum likelihood regression estimator in terms of the mean squared error. Analogous results for the ridge maximum likelihood estimators of covariance and precision matrix are presented.