Magazine
Estimation and Simulation for a Nonhomogeneous Poisson Process (303766)
*Lawrence Mark Leemis, The College of William & MaryKeywords: Intensity function estimation, process generation, stochastic processes.
We consider the problems of (1) estimating the intensity function of a nonhomogeneous Poisson process from event data and (2) simulating the event times in a nonhomogeous Poisson process with a given intensity function. The cases of raw data and count data are treated separately. In addition, the cases of a finite horizon (for example, arrivals to a bank during working hours) and infinite horizon (for example, arrivals to an emergency room) are treated separately. Inversion and thinning are presented as algorithms for generating events from a nonhomogeous Poisson process.