No label defined (Q3704764)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3932239
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3932239 |
Statements
1985
0 references
stopping time
0 references
sequential probability ratio tests
0 references
asymptotic behaviour of error probabilities
0 references
expected sample sizes
0 references
0.8278005719184875
0 references
The sequential probability ratio tests to distinguish between probability measures \(P_ 1\) and \(P_ 2\) are considered. Let \(f_ j=dP_ j/d\mu\) for \(j=1,2\), \(z_ i=\log (f_ 2(x_ i)^ 2/f_ 1(x_ i))\), \(S_ n=\sum^{n}_{i=1}z_ i\), \(a>0\), \(b>0\). It is assumed that \(E_ j| z_ 1| <\infty\) for \(j=1,2\) and \(P_ 1(z_ 1<y)\) has non- zero absolutely continuous component with respect to the Lebesgue measure. Denote \(N=\min \{n:\) \(S_ n\not\in (-a,b)\}.\) In this paper, based on a factorization of \(1-E_ je^{\lambda z_ 1}\) for \(j=1,2\), an analytical method is used to obtain the asymptotic behaviour of error probabilities \(\alpha_ 1(a,b)=P_ 1(S_ N\geq b)\), \(\alpha_ 2(a,b)=P_ 2(S_ N\leq -a)\) and expected sample sizes \(E_ jN\) for \(j=1,2\), when a,b\(\to \infty\).
0 references