Computational and Mathematical Methods in Medicine

Research Article

Regularized Weighted Nonparametric Likelihood Approach for High-Dimension Sparse Subdistribution Hazards Model for Competing Risk Data

Table 2

Results of the simulation studies for the Fine and Gray model with 5 informative variables () for scenario. Values shown are means (standard deviations) of each performance measure over 500 replicates (: ~40% average censoring).


		0.2			0.5			0.8
	Method	No. selected variables	TPR	FDR	No. selected variables	TPR	FDR	No. selected variables	TPR	FDR

200	ALASSO	35.702 (1.969)	0.779 (0.136)	0.006 (0.004)	36.054 (2.027)	0.900 (0.116)	0.006 (0.004)	30.261 (2.202)	0.953 (0.085)	0.005 (0.005)
	AENET	35.651 (1.766)	0.775 (0.163)	0.006 (0.003)	35.791 (1.745)	0.891 (0.124)	0.006 (0.004)	30.092 (1.584)	0.937 (0.101)	0.005 (0.003)
	SCAD	36.763 (3.429)	0.740 (0.166)	0.007 (0.006)	36.723 (2.717)	0.908 (0.108)	0.006 (0.006)	27.101 (2.831)	0.940 (0.101)	0.004 (0.003)
	SCAD-L₂	35.742 (2.458)	0.795 (0.148)	0.006 (0.005)	36.783 (3.299)	0.896 (0.120)	0.006 (0.005)	26.511 (2.062)	0.946 (0.094)	0.004 (0.003)
	MCP	26.234 (4.008)	0.690 (0.139)	0.004 (0.005)	25.870 (1.983)	0.847 (0.124)	0.004 (0.004)	24.384 (1.805)	0.922 (0.102)	0.004 (0.004)
	MCP-L₂	25.691 (2.124)	0.738 (0.133)	0.004 (0.005)	25.921 (2.146)	0.857 (0.137)	0.004 (0.004)	23.531 (2.654)	0.942 (0.099)	0.004 (0.004)
	Boosting (Binder)	40.524 (4.171)	0.966 (0.076)	0.007 (0.006)	39.447 (3.187)	0.994 (0.034)	0.007 (0.007)	38.602 (4.211)	0.990 (0.044)	0.007 (0.006)
	Oracle	5.000	1.000	0.000	5.000	1.000	0.000	5.000	1.000	0.000

400	ALASSO	37.795 (1.436)	0.942 (0.108)	0.007 (0.003)	35.522 (0.999)	0.990 (0.031)	0.006 (0.002)	29.581 (0.923)	1.000 (0.020)	0.005 (0.002)
	AENET	36.764 (1.534)	0.946 (0.093)	0.006 (0.003)	35.864 (1.137)	0.990 (0.044)	0.006 (0.003)	29.833 (1.234)	1.000 (0.000)	0.005 (0.002)
	SCAD	36.510 (1.956)	0.908 (0.117)	0.006 (0.004)	31.754 (1.774)	0.979 (0.061)	0.005 (0.003)	26.813 (1.523)	0.999 (0.034)	0.004 (0.003)
	SCAD-L₂	35.722 (1.590)	0.902 (0.113)	0.006 (0.003)	31.122 (1.297)	0.988 (0.048)	0.005 (0.003)	26.181 (1.296)	0.999 (0.034)	0.004 (0.003)
	MCP	25.553 (1.517)	0.874 (0.121)	0.004 (0.003)	22.701 (1.364)	0.963 (0.083)	0.004 (0.003)	20.455 (0.869)	0.992 (0.039)	0.003 (0.002)
	MCP-L₂	25.382 (1.388)	0.897 (0.115)	0.004 (0.003)	22.571 (0.987)	0.982 (0.057)	0.004 (0.003)	20.482 (1.020)	0.998 (0.020)	0.003 (0.002)
	Boosting (Binder)	39.452 (3.770)	0.994 (0.034)	0.007 (0.008)	37.752 (3.066)	1.000 (0.000)	0.006 (0.006)	38.330 (3.714)	1.00 (0.000)	0.007 (0.006)
	Oracle	5.000	1.000	0.000	5.000	1.000	0.000	5.000	1.000	0.000

TPR: true positive rate; FPR: false positive rate; : sample size.