P. Ramírez Cobo, R. E. Lillo Rodríguez, Y. G. Yera
The Batch Markovian Arrival Process (BMAP) is a general class of point processes for which the inter-event times are correlated and non-exponentially distributed. In addition, events can occur in batches which may also be correlated. BMAPs have been widely considered in the literature but mainly from a theoretical viewpoint. And less works have been devoted to study statistical inference which is of crucial importance in operational risk contexts, often characterized by dependent and simultaneous risk events. In this work, we consider the estimation for a wide subclass of BMAPs, namely, the Batch Markov-Modulated Poisson processes (BMMPP) which generalize the well-known Markov-Modulated Poisson process (MMPP). A matching moments technique, supported by a theoretical result that characterizes the process in terms of its moments, is considered. Numerical results with a real dataset related to operational risk will be presented to illustrate the performance of the novel approach.
Palabras clave / Keywords: loss modeling, batch Markovian arrival process, dependent losses times, point process, PH distribution, operational risk, value at risk
Programado
Sesión invitada SI01 Procesos Estocásticos: Modelización y Aplicaciones I (Organizadora: Inés Mª del Puerto García)
30 de mayo de 2018 10:50
Sala 2
