B0782
Title: Goodness-of-fit test for Cox processes and generalizations
Authors: Nuria Ruiz-Fuentes - University of Jaen (Spain) [presenting]
Paula Bouzas - University of Granada (Spain)
Abstract: Assessing whether a sample path of an observed counting process follows a Cox process model is a challenge, especially if any fixed structure of the intensity process is assumed. A goodness-of-fit test is derived for a Cox process in different situations, when the intensity of the null hypothesis is known or when it has to be estimated. Considering a compound Cox process, it is usual that the interest lies in the points with certain given marks. These points are proven to be a Cox process and the new intensity can be calculated or estimated using the representation theorem of the compound Cox process. Therefore, the goodness-of-fit test can be extended to this new process. The derived hypothesis tests are illustrated with several simulations of Cox and compound Cox processes. The simulated sample paths were tested directly and after disrupting them in different ways to check the performance of the test. The high percentage of correct decisions proves the accuracy of the proposed tests.
Journal