Conditional boundary crossing probabilities, with application to change- point problems (Q1103290)

Let \(X_ i\), \(i\geq 1\), be i.i.d. N(0,1) random variables, \(S_ n=X_ 1+...+X_ n\), \(U_ n=X^ 2_ 1+...+X^ 2_ n\). Given a twice differentiable function g(t), \(0<t\leq 1\), and \(m\geq 1\), let \[ \tau =\tau_ m=\inf \{n\leq m: S_ n\geq m g(n/m)\}. \] Denote \[ P^{(m)}_{x,y}(A)=P(A| S_ m=x,\quad U_ m=y)\quad and \] \[ \nu (t)=2t^{-2}\exp \{-2\sum^{\infty}_{n=1}n^{-1}\Phi (- tn^{1/2}/2)\}, \] where \(\Phi\) is the N(0,1) distribution function. The main result is an asymptotic expression for \(P^{(m)}_{\xi,\lambda}(\tau <m)\) (m\(\to \infty)\) given in terms of \(\nu\) and g. Here \(\xi\) and \(\lambda\) depend on m. The method extends that of \textit{D. Siegmund} [Ann. Probab. 10, 581-588 (1982; Zbl 0487.60028)]. For details on the function \(\nu\) see Ch. 10 of \textit{D. Siegmund}, Sequential analysis. Tests and confidence intervals. (1985; Zbl 0573.62071). The result is applied to a normal change-point problem to approximate null distributions of test statistics and to obtain approximate confidence sets for the change-point.