scientific article; zbMATH DE number 176299

From MaRDI portal
Publication:4035409





[For the entire collection see Zbl 0752.00047.] Assume that the distribution of a \(d\)-dimensional random point pattern (i.e. a point process \(\Psi\) in \(R^ d\)) depends on a vector \(\theta=(\theta_ 1,\dots,\theta_ s)\) of unknown parameters which lies in a compact subset \(\Theta\) of \(R^ s\). A minimum-contrast method in the sense of \textit{J. Pfanzagl} [Metrika 17, 82-91 (1971; Zbl 0216.478)] is considered for estimating the unknown parameter vector \(\theta\in\Theta\) from a single realization of \(\Psi\) in a sufficiently large sampling window \(A\subset R^ d\). It is assumed that \(\Psi\) is stationary and that the second moment \(E[\Psi(B)]^ 2\) of the number of points in \(B\) is finite for each bounded Borel set \(B\subset R^ d\). The contrast function is defined by a certain linear combination of first- order and second-order moment functions of \(\Psi\). Asymptotic properties of a sequence \((\widehat\theta_ n)\) of mc- estimates are studied corresponding to an unboundedly increasing sequence \((A_ n)\) of sampling windows. It is shown that, under some additional conditions, \((\widehat\theta_ n)\) is strongly consistent provided that \(\Psi\) is ergodic. For the case that \(\Psi\) is a Poisson cluster process, conditions are given for asymptotic normality of \((\widehat\theta_ n)\).











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 Q4035409)