日本統計学会誌
Online ISSN : 2189-1478
Print ISSN : 0389-5602
ISSN-L : 0389-5602
UNBIASED VARIABLE SELECTION PROCEDURE IN REGRESSION MODELS
Yasushi Nagata
著者情報
ジャーナル フリー

1987 年 17 巻 2 号 p. 175-183

詳細
抄録
The unbiasedness of the variable selection procedure in linear regression models is considered. Sawa and Takeuchi (1977) treated two model case, which is extended to three model case in this paper. Two kinds of definitions for the unbiased variable selection procedures are given: for the weakly unbiased one and for the strongly unbiased one. The weakly unbiased procedure based on Mallows' Cp statistic is obtained, and is compared with the original Cp statistic.
著者関連情報
© Japan Statistical Society
前の記事 次の記事
feedback
Top