pcoxtime: Penalized Cox Proportional Hazard Model for Time-dependent Covariates

2021/02/03 by Cygu, Steve, Dushoff, Jonathan, Bolker, Benjamin M.
#14J60 #Computation (stat.CO) #FOS: Computer and information sciences #Methodology (stat.ME)

paper · doi:10.48550/arxiv.2102.02297

Abstract

The penalized Cox proportional hazard model is a popular analytical approach for survival data with a large number of covariates. Such problems are especially challenging when covariates vary over follow-up time (i.e., the covariates are time-dependent). The standard R packages for fully penalized Cox models cannot currently incorporate time-dependent covariates. To address this gap, we implement a variant of gradient descent algorithm (proximal gradient descent) for fitting penalized Cox models. We apply our implementation to real and simulated data sets.

Related

• Survival of the fittest Cox model: Pivotal variable selection for time-to-event data
• <i>L</i>₁Penalized Estimation in the Cox Proportional Hazards Model
• Using fractional polynomials and restricted cubic splines to model non‐proportional hazards or time‐varying covariate effects in the Cox regression model
• Structured Learning in Time-dependent Cox Models
• A novel approach to estimate the Cox model with temporal covariates and its application to medical cost data