In this paper we obtain an asymptotic expansion for the distribution of Hotelling's T 2 -statistic T 2 under nonnormality when the sample size is large. In the derivation we find an explicit Edgeworth expansion of the multivariate t-statistic. Our method is to use the Edgeworth expansion and to expa
Asymptotic null and nonnull distribution of Hotelling's T2-statistic under the elliptical distribution
β Scribed by Toshiya Iwashita
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 743 KB
- Volume
- 61
- Category
- Article
- ISSN
- 0378-3758
No coin nor oath required. For personal study only.
β¦ Synopsis
This paper is concerned with asymptotic distribution of Hotelling's T 2-statistic under the elliptical distribution for the null hypothesis and the local alternative under the elliptical distribution. Asymptotic expansions for the distribution of T 2 for the null case and the local alternative are given up to the order N -t, where N is the sample size. The percentiles of T 2 and the approximate powers are calculated to evaluate the efl'ect of the elliptical distribution for some numerical examples.
Also to evaluate the effect of an adjustment of Bartlett type to Hotelling's T 2 tbr the local alternative, the approximate power of adjusted 7 ,2 is calculated in comparison with one of nonadjusted 7,2.
π SIMILAR VOLUMES
This paper is concerned with investigation into the behavior of the likelihood ratio test statistic G2 when the alternative hypothesis M ( Q ) is the true model. Exact moments of G2 are computed empirically and three approximations are considered for approximating the non-null distribution of G2. Ou
This paper is concerned with the null distribution of test statistic T for testing a linear hypothesis in a linear model without assuming normal errors. The test statistic includes typical ANOVA test statistics. It is known that the null distribution of T converges to w 2 when the sample size n is l
In this paper, an upper bound of van Zwet on the characteristic function of simple linear rank statistics is, under the null-hypothesis, generalized to the characteristic function of the exponentially tilted distribution. Our present result solves a di cult and crucial step in obtaining a rate of co