scientific article; zbMATH DE number 3896137

From MaRDI portal
Publication:3675366





Let \(x^ n=(x_ 1,..., x_ n)\) be an observed process with a distribution function \(P_{\theta}(x^ n)\). A general procedure for constructing a sequential confidence set for the parameter \(\theta\) is discussed in the first part of this paper. The second part is devoted to finding a sequential test for two composite hypotheses \(H_ 0: p(x)\in P_ 0\) and \(H_ 1: p(x)\in P_ 1\) where p(x) is the density function of the i.i.d. \(x_ j\), and \(P_ 0\), \(P_ 1\) are nonoverlapping families. It is shown that if \(P_ 0\) and \(P_ 1\) are convex, then the sequential test has a minimal average number of observations under the assumption of neglecting by overshoot effects.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3675366)