A fast adaptive Lasso for the cox regression via safe screening rules

Abstract

Some interesting recent studies have shown that safe feature elimination screening algorithms are useful alternatives in solving large scale and/or ultra-high-dimensional Lasso-type problems. However, to the best of our knowledge, the plausibility of adapting the safe feature elimination screening algorithm to survival models is rarely explored. In this study, we first derive the safe feature elimination screening rule for adaptive Lasso Cox model. Then, using both simulated and real-world datasets, we demonstrate that the resulting algorithm can outperform Lasso Cox and adaptive Lasso Cox prediction methods in terms of its predictive performance. In addition to its good predictive performance, we illustrate that the proposed algorithm has a key computational advantage over the above competing methods in terms of computation efficiency.

Keywords:

• Adaptive Lasso
• cox model
• high-dimensional survival data
• safe feature elimination screening

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was partly supported by National Social Science Foundation of China (17BTJ019) and supported by the Fundamental Research Funds for the Central Universities of Central South University (2020zzts361).