scientific article; zbMATH DE number 495596

From MaRDI portal
Publication:4277463





The paper presents results on the following detection problem: \[ H_ 0: \theta (t) \in \Theta_ 0 \subset \Theta \text{ vs. } H_ 1:\theta (t) \in \Theta_ 1 \subset \Theta \] when the observations are \[ X(t_ n, \omega_ 0)=\theta (t_ n)+B(t_ n, \omega_ 0),\;0 \leq t_ 1<t_ 2< \cdots < t_ n< \cdots \] with \(B\) a Cauchy process, and the paramter sets are disjoint. The main result says that ``uniform asymptotic separation relative to the observation times of \(H_ 0\) and \(H_ 1\) obtains as soon as \(\Theta_ 0\) and \(\Theta_ 1\) can be ``separated by two functions whose difference has successive increments which are not too different (in an appropriate sense) from \(t_ n-t_{n-1}\). The remainder of the paper is devoted to the study of the consequences of the main result which may be summarized as follows: The sensitivity of the law of \(\theta + B\) to variations in \(\theta\) is quite higher for \(B\) Cauchy than for \(B\) Gaussian. In particular, \(P_{\theta_ 0+B} \perp P_{\theta_ 1+B}\), as soon as \(\theta_ 0\) and \(\theta_ 1\) are ``slightly different.











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 Q4277463)