B1036
Title: Feature selection for competing risks models: A comparison
Authors: Marialuisa Restaino - University of Salerno (Italy) [presenting]
Abstract: In the analysis of time-to-event data, competing risks data are encountered when individuals may fail from multiple causes, and the occurrence of one failure event precludes the others from happening. To analyze the effects of covariates on the hazard function, both the cause-specific hazard (CSH) model and the subdistribution hazard (SDH) model. The main difference is in the definition of the risk set. In CSH, subjects who experience the competing events are treated as censored, while in SDH they are included in the risk set. In both settings, and in presence of a large number of variables, it becomes crucial to identify those variables that may affect the hazard. While in the CSH model, screening and variable selection methods developed for Cox model can be easily extended, for the SDH approach, naive applications of these procedures may be problematic and not suitable. It is due to the definition of the risk set. In the present work, the aim is to compare the performance of some existing methods for screening and selecting the most significant variables, for both CSH and SDH models, for highlighting their main advantages and disadvantages and proposing a new procedure able to identify the relevant covariates in the framework of high and ultra-high dimensions and also in presence of highly correlated variables.
