Nearly-Optimal Sequential Tests for Finitely Many Parameter Values
Lorden, Gary
Ann. Statist., Tome 5 (1977) no. 1, p. 1-21 / Harvested from Project Euclid
Combinations of one-sided sequential probability ratio tests (SPRT's) are shown to be "nearly optimal" for problems involving a finite number of possible underlying distributions. Subject to error probability constraints, expected sample sizes (or weighted averages of them) are minimized to within $o(1)$ asymptotically. For sequential decision problems, simple explicit procedures are proposed which "do exactly what a Bayes solution would do" with probability approaching one as the cost per observation, $c$, goes to zero. Exact computations for a binomial testing problem show that efficiencies of about 97${\tt\%}$ are obtained in some "small-sample" cases.
Publié le : 1977-01-14
Classification:  Sequential probability ratio test,  Bayes solution,  asymptotic optimality,  62L10,  62F20
@article{1176343737,
     author = {Lorden, Gary},
     title = {Nearly-Optimal Sequential Tests for Finitely Many Parameter Values},
     journal = {Ann. Statist.},
     volume = {5},
     number = {1},
     year = {1977},
     pages = { 1-21},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1176343737}
}
Lorden, Gary. Nearly-Optimal Sequential Tests for Finitely Many Parameter Values. Ann. Statist., Tome 5 (1977) no. 1, pp.  1-21. http://gdmltest.u-ga.fr/item/1176343737/