M. GONZALEZ BERNAL, R. E. Lillo, P. Ramirez Cobo
This research uses a real dataset of emergency calls, categorized into five types by hospital priority. The dataset's monthly counts reveal significant correlations, suggesting a modeling approach using an extension of the bivariate Markov Modulated Poisson Process with two states, as introduced by Yera et al. [2021]. Unlike Yera et al. [2021], which included train travel distance and time between failures, the hospital dataset only provides counting process information. Previous literature on inference with counting data, such as Nasr et al. [2018], Gonzalez et al. [2024], and Okamura et al. [2009], will be considered. The research adapts the bivariate process from Yera et al. [2021] to a five-variate version of the Marshall-Olkin exponential distribution, as described by Marshall and Olkin [1967]. The proposed solution involves the smallest number of parameters possible to deduce properties of interest for inferring the new process.
Keywords: Markov Modulated Poisson Process, Counting processes, Marshall-Olkin exponential distribution.
Scheduled
Stochastic processes and their applications I
June 13, 2025 11:00 AM
Auditorio 2. Leandre Cristòfol