Biomarker detection using corrected degree of domesticity in hybrid social network feature selection for improving classifier performance

Table 4 Results of the simulated data with rho = 0.60

FS	CD	Acc.	Sens.	Spe.	AUC
SVM SQRT-RFE+CS	Louvain	0.771	0.810	0.732	0.847
SVM SQRT-RFE+IG		0.769	0.828	0.710	0.845
SVM SQRT-RFE+RF		0.821	0.872	0.770	0.907
IG+CS		0.781	0.798	0.764	0.863
IG+RF		0.687	0.756	0.618	0.771
CS+RF		0.689	0.852	0.526	0.789
SVM SQRT-RFE+CS	Walktrap	0.740	0.676	0.804	0.811
SVM SQRT-RFE+IG		0.769	0.828	0.710	0.845
SVM SQRT-RFE+RF		0.821	0.872	0.770	0.907
IG+CS		0.737	0.682	0.792	0.829
IG+RF		0.687	0.756	0.618	0.771
CS+RF		0.689	0.852	0.526	0.789
SVM SQRT-RFE+CS	Infomap	0.771	0.810	0.732	0.847
SVM SQRT-RFE+IG		0.769	0.828	0.710	0.845
SVM SQRT-RFE+RF		0.821	0.872	0.770	0.907
IG+CS		0.781	0.798	0.764	0.863
IG+RF		0.687	0.756	0.618	0.771
CS+RF		0.689	0.852	0.526	0.789

FS: feature selection method used in the first step of the algorithm; CD: community detection method used in the third step of the algorithm; Acc.: accuracy; Sens: sensitivity; Spe.: specificity; AUC: area under the ROC curve

ISSN: 1471-2105