Two-dimensional Sequence Homogeneity Testing Against Mixture Alternative
Abstract
The behavior of linear rank statistics is investigated on models in which various subsequences of observations follow different statistical distributions. Such data can be interpreted both as models of a finite number distribution mixtures and as dependence models. We apply data set simulation to obtain estimates of average and variance of used rank statistics. The modeled and asymptotic results are enough close.
References
Lausen B., Shumacher H., " Maximally selected rank statistics", Biometrics, 48, pp. 73-85, 1 992.
Haroutunian E., Safaryan I. "Distributions mixture division with a stratifying parameter", submitted for publication.
Safaryan I., Haroutunian E. "A Common approach to the distributions mixture identification and dependence models analysis", Proceedings of CSIT 2003, pp. 1 84-1 86.
Haroutunian E. and Safaryan I. "Nonparametric consistent estimation of the change moment of random sequence properties", Transactions of Institute for Informatics and Automation Problems of NAS of RA and of YSU, Mathematical P roblems of Computer Science, vol. 17, pp. 76-85, 1997.
Hothorn T., Lausen B., "On the exact distribution of maximally selected rank statistics", Comp. Statist. and Data Anal., vol. 43, pp. 121- 137, 2003.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
