Abstract:

Count time series are frequently encountered in biomedical, epidemiological, and public health applications. In principle, such series may exhibit three distinctive features: overdispersion, zero-inflation, and temporal correlation. Devising modeling frameworks that are sufficiently general to accommodate all three of these characteristics poses a daunting challenge. To address this challenge, we discuss the development of frameworks based on both observation-driven and parameter-driven modeling formulations. We illustrate the latter development by presenting a flexible class of dynamic models in the state-space framework. For parameter estimation, we derive a Monte Carlo Expectation-Maximization (MCEM) algorithm, where particle filtering and particle smoothing methods are employed to approximate the high-dimensional integrals in the E-step of the algorithm. To exemplify the proposed methodology, we consider an application based on the evaluation of a participatory ergonomics intervention, which is designed to reduce the incidence of workplace injuries among a group of hospital cleaners. The data consists of aggregated monthly counts of work-related injuries that were reported before and after the intervention. (This is joint work with Gideon Zamba, Ming Yang, and Fan Tang.)