Iterative L1/2 Regularization Algorithm for Variable Selection in the Cox Proportional Hazards Model
In this paper, we investigate to use theL 1/2 regularization method for variable selection based on the Cox’s proportional hazards model. The L 1/2 regularization method isa reweighed iterative algorithm with the adaptively weighted L 1 penalty on regression coefficients. The algorithm of theL 1/2 regularization method can be easily obtained by a series of L 1 penalties. Simulation results based on standard artificial data show that theL 1/2 regularization method can be more accurate for variable selection than Lasso and adaptive Lasso methods. The results from Primary Biliary Cirrhosis (PBC) dataset indicate theL 1/2 regularization method performs competitively.
KeywordsLasso L1/2 regularization Variable selection Cox model
Unable to display preview. Download preview PDF.
- 7.Xu, Z.B., Zhang, H., Wang, Y., Chang, X.Y.: L1/2 regularization. Science in China, Series F 40(3), 1–11 (2010)Google Scholar
- 8.Zhang, H., Xu, Z.B., Chang, X.Y., Liang, Y.: Variable selection and sparse reconstruction via L1/2 Regularization, Sparsity in Machine Learning and Statistic, London (2009)Google Scholar
- 9.Fu, W.: Penalized regression: the bridge versus the lasso. J. Comp. Graph. Statist. 7, 397–416 (1998)Google Scholar
- 10.Qian, J., Li, B., Chen, P.: Generating Survival Data in the Simulation Studies of Cox Model. In: 2010 Third International Conference on Information and Computing (ICIC), pp. 93–96 (2010)Google Scholar