## Transformations for multivariate statistics.(English)Zbl 1071.62049

Summary: This paper derives transformations for multivariate statistics that eliminate asymptotic skewness, extending results of N. Niki and S. Konishi [Ann. Inst. Stat. Math. 38, 371–383 (1986; Zbl 0609.62075)]. Within the context of valid Edgeworth expansions for such statistics we first derive the set of equations that such a transformations must satisfy and second propose a local solution that is sufficient up to the desired order.
Application of these results yields two useful corollaries. First, it is possible to eliminate the first correction term in an Edgeworth expansion, thereby accelerating convergence to the leading term normal approximation. Second, bootstrapping the transformed statistic can yield the same rate of convergence of the double, or prepivoted, bootstrap of R. Beran [J. Am. Stat. Assoc. 83, No. 403, 687–697 (1988; Zbl 0662.62024)], applied to the original statistic, implying a significant computational saving.
The analytic results are illustrated by application to the family of exponential models, in which the transformation is seen to depend only upon the properties of the likelihood. The numerical properties are examined within a class of nonlinear regression models (logit, probit, Poisson, and exponential regressions), where the adequacy of the limiting normal and of the bootstrap [utilizing the $$k$$-step procedure of D. W. K. Andrews, Econometrica 70, 119–162 (2002)] as distributional approximations is assessed.

### MSC:

 62H05 Characterization and structure theory for multivariate probability distributions; copulas 65C60 Computational problems in statistics (MSC2010) 62B05 Sufficient statistics and fields

### Citations:

Zbl 0609.62075; Zbl 0662.62024
Full Text:

### References:

 [1] DOI: 10.2307/1912315 · Zbl 0373.62065 [2] DOI: 10.2307/1911222 · Zbl 0349.62070 [3] DOI: 10.1007/BF02482524 · Zbl 0609.62075 [4] DOI: 10.1093/biomet/67.2.311 · Zbl 0436.62020 [5] DOI: 10.1111/1468-2354.00026 [6] DOI: 10.1214/aoms/1177728652 · Zbl 0058.35404 [7] DOI: 10.2307/2938228 · Zbl 0729.62013 [8] DOI: 10.1214/aos/1176344134 · Zbl 0396.62010 [9] DOI: 10.2307/2289292 · Zbl 0662.62024 [10] DOI: 10.1111/1468-0262.00086 · Zbl 1056.62506 [11] DOI: 10.1111/1468-0262.00271 · Zbl 1104.62315 [12] van Garderen, Econometric Theory 15 pp 771– (1997) [13] DOI: 10.1093/biomet/71.2.233 · Zbl 0573.62026 [14] DOI: 10.1016/S0167-7152(00)00209-1 · Zbl 0988.62009 [15] Linton, Econometric Theory 13 pp 558– (1997) [16] Hougaard, Journal of the Royal Statistical Society, Series B 44 pp 244– (1982) [17] DOI: 10.1016/S0304-4076(99)00042-1 · Zbl 0971.62013 [18] DOI: 10.1016/0304-4076(94)90092-2 · Zbl 04518855 [19] DOI: 10.2307/2171846 · Zbl 0888.62088 [20] Skovgaard, Scandinavian Journal of Statistics 8 pp 207– (1981) [21] DOI: 10.2307/1910999 · Zbl 0557.62018 [22] DOI: 10.2307/1914165 · Zbl 0595.62119 [23] DOI: 10.2307/1913972 · Zbl 0338.62063 [24] DOI: 10.2307/1911222 · Zbl 0349.62070
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.