条件付指数型分布族の一様最強力不偏検定とその応用   :   <Article>Uniformly Most Powerful Unbiased Test for Conditionally Exponential Family Distribution and Its Applications : 論説 

作成者 大森, 裕浩
作成者 (ヨミ) オオモリ, ヤスヒロ
作成者の別表記 OMORI, Yasuhiro
日本十進分類法 (NDC) 331.19, 417
内容 Lehmann and Sheffe (1955) derived a uniformly most powerful test for an exponential family distribution. This assumption of exponentiality, however, is not always satisfied. For example, Poisson distribution has been frequently used to analyze discrete data. It assumes that its mean is equal to its variance, but data often suggest that the variance is larger than the mean. To explain this overdispersion, a parameter is considered random so that the distribution belongs to a conditionally exponential family distribution (Lawless (1987a), Lawless (1987b), Thall (1988), Dean and Lawless (1989)). This paper derives a uniformly most powerful unbiased test for a conditionally exponential family distribution without specifying any distribution for a conditioning random variable and gives its applications.
公開者 千葉大学経済学会
コンテンツの種類 紀要論文 Departmental Bulletin Paper
DCMI資源タイプ text
ファイル形式 application/pdf
ISSN 0912-7216
NCID AN10005358
掲載誌情報 千葉大学経済研究 Vol.8 no.4 page.169-183 (19940302)
情報源 Economic journal of Chiba University
言語 日本語
著者版フラグ publisher

Total Access Count:

927 times.

Related Materials in