Equivariant estimation of quantile vector of two normal populations with a common mean

Year 2019, Volume: 48 Issue: 1, 255 - 273, 01.02.2019

Manas Ranjan Tripathy , Adarsha Kumar Jena Somesh Kumar

Abstract

The problem of estimating quantile vector $\theta=(\theta_1,\theta_2)$ of two normal populations, under the assumption that the means ($\mu_i$s) are equal has been considered. Here $\theta_i=\mu+\eta\sigma_i,$ $i=1,2,$ denotes the $p^{th}$ quantile of the $i^{th}$ population, where $\eta=\Phi^{-1}(p)$, $0<p<1,$ and $\Phi$ denotes the c.d.f. of a standard normal random variable. The loss function is taken as sum of the quadratic losses. First, a general result has been proved which helps in constructing some improved estimators for the quantile vector $\theta.$ Further, classes of equivariant estimators have been proposed and sufficient conditions for improving estimators in these classes are derived. In the process, two complete class results have been proved. A numerical comparison of these estimators are done and recommendations have been made for the use of these estimators. Finally, we conclude our results with some practical examples.

Keywords

Equivariant estimator, Estimation of quantiles, Complete class results, Common mean, Inadmissibility, Relative risk comparison

References

• Alexander, T. L. and Chandrasekar, B. Simultaneous equivariant estimation of the parameters of matrix scale and matrix location-scale models, Statistical Papers 46, 483-507, 2005.
• Bai, S. K. and Durairajan, T. M. Simultaneous equivariant estimation of the parameters of linear models, Statistical Papers 39, 125-134, 1998.
• Brewster, J. F. and Zidek, J. V. Improving on equivariant estimators, Annals of Statistics 2, 21-38, 1974.
• Brown, L. D. and Cohen, A. Point and confidence estimation of a common mean and recovery of interblock information, Annals of Statistics 2 (5), 963-976, 1974.
• Chang, C. H. and Pal, N. Testing on the common mean of several normal distributions, Computational Statistics and Data Analysis 53, 321-333, 2008.
• Hines, W. W., Montgomery, D. C., Goldsman, D. M. and Borror, D. M. Probability and Statistics in Engineering, John Wiley, New Work, 2008.
• Keating, J. P. and Tripathi, R. C. Percentiles, estimation of 'Encyclopedia of Statistical Sciences, VI, 668-674, 1985.
• Kiefer, J. Invariance minimax sequential estimation and contineous time processes, Annals of Mathematical Statistics 28, 573-601, 1957.
• Kumar, S. and Tripathy, M. R. Estimating quantiles of normal populations with a common mean, Communications in Statistics-Theory and Methods 26, 115-118, 2011.
• Lin, S. H. and Lee, J. C. Generalized inferences on the common mean of several normal populations, Journal of Statistical Planning and Inferences 134, 568-582, 2005.
• Moore, B. and Krishnamoorthy, K. Combining independent normal sample means by weighting with their standard errors, Journal of Statistical Computation and Simulations 58, 145-153, 1997.
• Pal, N., Lin, J. J., Chang, C. H. and Kumar, S. A revisit to the common mean problem: Comparing the maximum likelihood estimator with the Graybill-Deal estimator, Computational Statistics and Data Analysis 51, 5673-5681, 2007.
• Rohatgi, V. K. and Saleh, A. Md. E. An Introduction to Probability and Statistics, John Wiley, 2nd Edition, New Work, 2003.
• Rukhin, A. L. A class of minimax estimators of a normal quantile, Statistics and Probability Letters 1, 217-221, 1983.
• Rukhin, A. L. Admissibility and minimaxity results in the estimation problem of exponential quantiles, Annals of Statistics 14, 220-237, 1986.
• Saleh, A. K. Md. E. Estimating quantiles of exponential distributions, In:Statistics and Related Topics, Csorgo, M., Dawson, D., Rao, J. N. K., Saleh, A. K. Md. E. (Eds.), North Holland, Amsterdam, 279-283, 1981.
• Sharma, D. and Kumar, S. Estimating quantiles of exponential populations, Statistics and Decisions, 12, 343-352, 1994.
• Tripathy, M. R. and Kumar, S. Estimating a common mean of two normal populations, Journal of Statistical Theory and Applications 9 (2), 197-215, 2010.
• Tsukuma, T. Simultaneous estimation of restricted location parameters based on permutation and sign-change, Statistical Papers 53, 915-934, 2012.
• Vazquez, G., Duval, S., Jacobs Jr, D. R. and Silventoinen, K. Comparison of body mass index, waist cicumference and waist/hip ratio in predicting incident diabetes: A meta analysis, Epidemilogic Reviews 29 (1), 115-128, 2007.
• Zidek, J. V. Inadmissibility of the best invariant estimators of extreme quantiles of the normal law under squared error loss, Annals of Mathematical Statistics 40 (5), 1801-1808, 1969.
• Zidek, J. V. Inadmissibility of a class of estimators of a normal quantile, The Annals of Mathematical Statistics 42 (4), 1444-1447, 1971.