论文信息 - Restrictive adaptive tests for the treatment of the two-sample scale problem

Restrictive adaptive tests for the treatment of the two-sample scale problem

Abstract The two-sample scale problem is studied using special restrictive adaptive methods. A two-stage procedure is proposed. The first stage is based on order statistics and the second stage is based on rank statistics. We use transformations of order statistics which are dependent on measures of tailweight and skewness of distributions. These measures are estimated by quantiles. Adaptive tests are derived on the basis of the asymptotic relative efficiences of the rank tests. It is shown by simulation experiments that these methods are also suitable for moderate sample sizes.

Wolfgang Kössler

[1] Representations of the space of distributions useful in robust estimation of location , 1982 .

[2] Richard A. Groeneveld,et al. Measuring Skewness and Kurtosis , 1984 .

[3] Anthony Ralston,et al. Mathematische Methoden für Digitalrechner , 1967 .

[4] B. S. Duran. A survey of nonparametric tests for scale , 1976 .

[5] On the asymptotic behaviour of selector statistics , 1988 .

[6] Herbert Büning. Robuste und adaptive Tests , 1991 .

[7] P. Sen,et al. Theory of rank tests , 1969 .

[8] Herbert Büning. Adaptive verteilungsfreie Tests , 1983 .

[9] R. Hogg. On adaptive statistical inferences , 1982 .

[10] H. Büning,et al. Nichtparametrische Statistische Methoden , 1994 .

[11] R. Hogg. Adaptive Robust Procedures: A Partial Review and Some Suggestions for Future Applications and Theory , 1974 .

[12] Robustness of Two-Sample Tests for Variances , 1984 .

[13] R. Serfling. Approximation Theorems of Mathematical Statistics , 1980 .

[14] Bruce W. Schmeiser,et al. An approximate method for generating symmetric random variables , 1972, CACM.

[15] Robert V. Hogg,et al. A Two-Sample Adaptive Distribution-Free Test , 1975 .