Estimation of Quantiles of Exponential Distributions with Minimum Error in Predicted Distribution Functions
Journal of the American Statistical Association · 1977 · Volume 72 · Issue 357
DOI: 10.1080/01621459.1977.10479931
Abstract
We consider the problem of estimating the quantiles of a one-parameter exponential distribution. Using the estimated quantiles to predict the distribution function, and attempting to minimize mean square errors for these predicted probabilities rather than the quantiles themselves, we are led to estimators differing from the natural ones obtained by appropriate scaling of the estimated mean of the distribution. The form of these estimators is studied for complete and censored samples, and also for the optimal single order statistic. Numerical comparisons are made, leading to the conclusion that for moderate sample sizes the estimators studied here are considerably more efficient than the natural ones, particularly for high quantiles.
Enable JavaScript for citation tools and additional paper interactions.