High-dimensional variable screening for survival endpoints

Speaker: Axel Benner, Senior Researcher and Group Leader "Statistics for Translational Oncology", Department of Biostatistics, German Cancer Research Center, Heidelberg, Germany,  


When performing survival analysis in very high dimensions, it is often required
to reduce the number of covariates using preliminary screening. During the last
years, a large number of variable screening methods for the survival context
have been developed. However guidance is missing which method to apply.
In this work, the literature on variable screening for survival is reviewed. A
novel screening procedure based on the distance correlation between martin-
gale residuals and covariates is proposed. The novel method and a selection of
existing methods are compared in simulations, involving both monotone and
non-monotone associations. Recommendations for the use of variable screen-
ing in practice are developed.

Published Aug. 12, 2018 3:06 PM - Last modified Nov. 20, 2018 2:00 PM