Inference for earthquake models: A self-correcting model (Q796230)

The paper studies asymptotic inference problems for point process models in which the conditional intensity function can be represented in the form E(dN(t) \(| H_{0,t})=\exp \{\alpha +\beta(t-\rho N(t)\}\), where \(H_{0,t}\) is the \(\sigma\)-field generated by the process in the time interval \(0\leq s<t\), \(\alpha\),\(\beta\),\(\rho\) are constants, and N(t) is the number of events that have occurred since the beginning of observations at time 0. Parameter estimates are shown to be consistent and, except under the null hypothesis of a Poisson process, normally distributed. Under the null hypothesis, however, the Hessian matrix is not asymptotically constant and the limiting distribution of the likelihood ratio statistic is not \(\chi^ 2\), but has a form related to that of the Cramér-von Mises \(\omega^ 2\) statistic for the goodness of fit test.