Lorden, Gary (1976) 2SPRT'S and the modified KieferWeiss problem of minimizing an expected sample size. Annals of Statistics, 4 (2). pp. 281291. ISSN 00905364. http://resolver.caltech.edu/CaltechAUTHORS:LORas76

PDF
See Usage Policy. 942Kb 
Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechAUTHORS:LORas76
Abstract
A simple combination of onesided sequential probability ratio tests, called a 2SPRT, is shown to approximately minimize the expected sample size at a given point θ0 among all tests with error probabilities controlled at two other points, θ1 and θ2. In the symmetric normal and binomial testing problems, this result applies directly to the KieferWeiss problem of minimizing the maximum over θ of the expected sample size. Extensive computer calculations for the normal case indicate that 2SPRT's have efficiencies greater than 99% regardless of the size of the error probabilities. Accurate approximations to the error probabilities and expected sample sizes of these tests are given.
Item Type:  Article 

Additional Information:  © 1976 Institute of Mathematical Statistics. Received June 1974; revised July 1975. Research supported by the National Science Foundation under Grant GP37819. The author is indebted to F. R. Maiocco of the Jet Propulsion Laboratory for his very generous help with the computer calculations. The author also wishes to thank David Siegmund, Tze Lai, and Michael Klass for helpful discussions. 
Subject Keywords:  Sequential probability ratio test; Bayes solution; asymptotic optimality 
Record Number:  CaltechAUTHORS:LORas76 
Persistent URL:  http://resolver.caltech.edu/CaltechAUTHORS:LORas76 
Alternative URL:  http://links.jstor.org/sici?sici=00905364%28197603%294%3A2%3C281%3A2ATMKP%3E2.0.CO%3B28 
Usage Policy:  No commercial reproduction, distribution, display or performance rights in this work are provided. 
ID Code:  5258 
Collection:  CaltechAUTHORS 
Deposited By:  Tony Diaz 
Deposited On:  06 Oct 2006 
Last Modified:  26 Dec 2012 09:04 
Repository Staff Only: item control page