A1298
Title: Penalized piecewise exponential distributional regression model for survival analysis
Authors: Jack Moore - University of Limerick (Ireland) [presenting]
Shirin Moghaddam - University of Limerick (Ireland)
Kevin Burke - University of Limerick (Ireland)
Abstract: The field of survival analysis is concerned with the modelling of time-to-event data, a form of data that arises in various application areas. A key application area of survival analysis is medical research, where interest lies in survival times of patients. Traditional parametric modelling approaches rely on distributions such as the Weibull, which can impose strong assumptions on the data at hand. In contrast, the piecewise exponential model offers a much more general parametric modelling approach: It has the capability of approximating any survival distribution, without prior knowledge of the underlying distribution. This is achieved by partitioning the time scale into intervals, within which the hazard rate is constant. Despite the versatility of the piecewise exponential model, it has historically been underutilized within the literature. This can perhaps be explained by the popularity of the Cox model. However, in recent years, there has been renewed interest in the piecewise exponential model, with various developments aimed at enhancing its viability. The purpose is to introduce the piecewise exponential model and present extensions aimed at improving this model. Specifically, a distributional regression structure is used. Thus, the framework enables flexible modelling of both the underlying baseline hazard and the nature of covariate effects, where the intervals/ pieces and covariates of the model are selected automatically via an adaptive lasso penalization.
