Univariate Feature Selection and Compound Covariate for Predicting Survival, Including Copula-Based Analyses for Dependent Censoring

Univariate feature selection and compound covariate methods under the Cox model with high-dimensional features (e.g., gene expressions). Available are survival data for non-small-cell lung cancer patients with gene expressions (Chen et al 2007 New Engl J Med) , statistical methods in Emura et al (2012 PLoS ONE) , Emura & Chen (2016 Stat Methods Med Res) , and Emura et al (2019). Algorithms for generating correlated gene expressions are also available. Estimation of survival functions via copula-graphic (CG) estimators is also implemented, which is useful for sensitivity analyses under dependent censoring (Yeh et al 2023 Biomedicines) and factorial survival analyses (Emura et al 2024 Stat Methods Med Res) .