Bayesian analysis of self-exciting point processes

Thời gian: 09:00 đến 10:00 Ngày 09/05/2016

Địa điểm: C2-714

Báo cáo viên: Evans Gouno

Tóm tắt:

This talk gives an overview of some bayesian techniques for statistical analysis of stochastic processes belonging to the family of self-exciting point processes (SEPP). A SEPP is a stochastic process with an intensity which depends not only on time but on the numbers and dates of the previous jumps of the process. Thus the intensity is itself a random process. After describing the main properties of SEPP, we consider the problem of estimating the parameters of the intensity. The classical maximum likelihood is presented before the Bayesian approach is investigated. We will illustrate the presentation with examples in different field such as seismology, neurology and reliability.