Non-parametric Quantile Estimation Through Stochastic Approximation.
Abstract
The extreme values which a random variable X may take on are usually best characterized by the quantiles of the random variable. Known non-parametric methods for the statistical estimation of extreme quantiles all suffer from serious shortcomings, however. In this thesis a robust and efficient method for quantile estimation is described; both the asymptotic and finite sample properties of the estimator are determined and computer implementations are given. Possible applications for the technique include the analysis of computer simulations and data analysis in large data bases or real time computer systems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1975
- Accession Number
- ADA014628
Entities
People
- David Walter Robinson
Organizations
- Naval Postgraduate School