No label defined (Q4720071)

The author studies statistical inference for marked point processes (MPP). A marked point process \(\{(\tau_k, \zeta_k), k\geq 1\}\) is a stochastic process where \(\tau_k\) denotes the time of occurrence of the \(k\)-th event \(\zeta_k\) which is realized in a certain mark space \(E\). A framework is established for stochastic models for MPP with an arbitrary mark space. Suppose the probability measure determining the MPP is governed by an unknown \(\vartheta \in\Theta\). The problem of estimation of \(\vartheta\) is studied. The asymptotic properties of the estimator of \(\vartheta\) are studied either by observing an increasing number of individuals during a fixed time period as \(n\to\infty\) or investigating a finite set of objects for time \(t\to\infty\). The unknown \(\vartheta\) can be a \(d\)-dimensional parameter \((\vartheta, \dots, \vartheta_d) \in \Theta\) contained in \(R^d\) or \(\vartheta= (\alpha,\beta)\) where \(\alpha\) is a function and \(\beta\in B\) contained in \(R^d\). Chapter 1 contains an introduction to the basic topics starting from the definition of a stochastic process, properties of continuous-time martingales and stochastic Stieltjes integrals, compensators and point processes. The theory of marked point processes with an arbitrary mark space is developed in Chapter 2. Statistical inference for such processes is investigated in Chapters 3 and 4. The book contains results of a Ph.D. thesis submitted by the author. It is well written and a welcome addition to the subject of statistical inference for stochastic processes.