A dependent and censored first hitting-time model with compound Poisson processes
Abstract
We consider a bivariate first hitting-time model in which durations are the crossing times of dependent compound Poisson processes with fixed thresholds. The identifiability of the model is discussed, and likelihood estimators of the model parameters are proposed. We obtain the asymptotic properties of the estimators and underline their finite sample performance with a simulation study on synthetic data. The practical applicability of our approach is demonstrated by an application using data from patients suffering from mushroom poisoning.
Fields of Study & Subjects
Cold Atom Physics and Bose-Einstein Condensates
Stochastic processes and statistical mechanics
Random Matrices and Applications
Publication details
- Journal / source
- Lifetime Data Analysis
- Volume
- 32
- Issue
- 2
- Publication date
- 27 April 2026
- Language
- EN