作成者 |
|
|
|
本文言語 |
|
出版者 |
|
|
発行日 |
|
収録物名 |
|
巻 |
|
号 |
|
開始ページ |
|
終了ページ |
|
出版タイプ |
|
アクセス権 |
|
Crossref DOI |
|
関連DOI |
|
|
関連URI |
|
|
関連情報 |
|
|
概要 |
Data analysed in medical a nd biomedical research are often found to have skewed distributions, so transformations are applied to the data in order to render them more tractable by ordinary normal the...ory analysis. The power-normal distribution, proposed by Goto, Uesaka and Inoue [4], is a distribution which specifies original observations before Box-Cox transformation. In this paper, we mention the maximum likelihood method to estimate the parameters of the power-normal distribution, and consider the sampling behaviours of the maximum likelihood estimates. And we show, through simulation experiment, that biases of the transforming parameter, resulted from neglecting the probability proportional constant term, do not seem to cause any troubles in case the aim is to attain approximate normality of the transformed observations, but they do have considerable infuluences on the estimates of the parmeters of near-normal distribution after the tranformation.続きを見る
|