Adaptive tests for stochastic processes in the ergodic case (Q1208955)

The hypothesis testing problem for two processes on a growing time interval when the distributions are known only up to a nuissance parameter \(\theta\) is considered. A test which is asymptotically efficient for a specified pair of null hypotheses and alternatives is called adaptive. A necessary and sufficient adaptation condition in the ergodic case is obtained. Several examples (concerning Gaussian processes in discrete and continuous time, and autoregressive processes) with explicitly constructed adaptive tests are studied.