Showing papers on "Sequential probability ratio test published in 1971"
TL;DR: In this paper, the PSPRT was compared to the Wald's SPRT with the same error probabilities for some parameter values, such as the Koopman-Darmois parameter.
Abstract: In testing a normal mean with known variance, or a Koopman-Darmois parameter, an initial fixed number n of observations is followed by Wald's SPRT procedure. The conditional SPRT, given, and optimality properties in a certain class of tests are noted. For some parameter values, the PSPRT may have a lower ASN than a Wald SPRT with the same error probabilities.
10 citations
TL;DR: A sequential test for correlation coefficients is presented in this article, which consists of reducing the test of a correlation coefficient to that of a variance ratio and making use of Helmert transformations of sequential observations.
Abstract: A sequential test for correlation coefficients is presented The method consists of reducing the test of a correlation coefficient to that of a variance ratio and making use of Helmert transformations of sequential observations The resulting test is identified as the sequential probability ratio test regarding the parameter of the Cauchy distribution The average sample number of the test is given and it is shown that the proposed sequential test yields an appreciable saving in the required average number of observations compared to the usual fixed sample size test
6 citations
TL;DR: In this paper, the authors investigated the distribution of sample size required to terminate a sequential test of a hypothesis on the single parameter, p, of the binomial distribution, and compared a non-truncated sequential test with a corresponding test of fixed sample size, N, in which the null hypothesis is accepted if there are at least ir, successes out of N trials.
Abstract: In a recent paper [l], we have investigated the distribution of sample size (no. of test stages) required to terminate a sequential test of a hypothesis on the single parameter, p, of the binomial distribution. It is well known that, for certain values of the true (unknown) p, the sequential test will require an unacceptably large number of test stages to terminate; therefore, any practical sequential test procedure must include a truncation rule. In [l], we compared a non-truncated sequential test with a corresponding test of fixed sample size, N, in which the null hypothesis is accepted if there are at least ir, successes out of N trials. For the null hypothesis Ho : p = p, , against the alternative
6 citations