GARS: Genetic Algorithm for the identification of a Robust Subset of features in high-dimensional datasets

BMC Bioinformatics

Table 4 Performance evaluation, testing FS methods on reduced ‘multi-class high-dimension’ datasets (1000 features)

		ACC	SEN	SPE	PPV	NPV	Time	Nfeats
3	GARS	0.92	0.90	0.95	0.88	0.94	26 min	18
3	RFE	0.94	0.92	0.96	0.91	0.96	3 s	999
3	SBF	0.95	0.94	0.97	0.94	0.97	1 min 58 s	289
3	rfGA	0.93	0.91	0.95	0.91	0.95	19 h 55 min	598
3	svmGA	–	–	–	–	–	–	–
3	LASSO	0.95	0.93	0.96	0.93	0.97	2 s	54
5	GARS	0.93	0.89	0.97	0.89	0.97	1 h 22 min	17
5	RFE	0.93	0.89	0.97	0.88	0.97	6 s	21
5	SBF	0.93	0.89	0.97	0.87	0.97	9 min 38 s	890
5	rfGA	–	–	–	–	–	–	–
5	svmGA	–	–	–	–	–	–	–
5	LASSO	0.96	0.93	0.98	0.93	0.98	2 s	74
7	GARS	0.90	0.84	0.97	0.81	0.97	3 h 6 min	16
7	RFE	0.95	0.91	0.98	0.89	0.98	13 s	999
7	SBF	0.95	0.92	0.99	0.90	0.99	16 min 7 s	959
7	rfGA	–	–	–	–	–	–	–
7	svmGA	–	–	–	–	–	–	–
7	LASSO	0.96	0.93	0.99	0.90	0.99	4 s	105
9	GARS	0.92	0.86	0.98	0.85	0.98	6 h 6 min	22
9	RFE	0.93	0.89	0.98	0.87	0.98	11 s	25
9	SBF	0.95	0.91	0.99	0.89	0.99	22 min 47 s	963
9	rfGA	–	–	–	–	–	–	–
9	svmGA	–	–	–	–	–	–	–
9	LASSO	0.96	0.92	0.99	0.90	0.99	6 s	123
11	GARS	0.93	0.88	0.99	0.86	0.98	10 h 31 min	19
11	RFE	0.94	0.90	0.99	0.88	0.99	17 s	999
11	SBF	0.95	0.91	0.99	0.89	0.99	30 min 44 s	976
11	rfGA	–	–	–	–	–	–	–
11	svmGA	–	–	–	–	–	–	–
11	LASSO	0.96	0.92	0.99	0.90	0.99	9 s	134

ACC Accuracy, SEN Sensitivity, SPE Specificity, PPV Positive Predictive Value, NPV Negative Predictive Value, AUC Area Under ROC Curve, Time average learning time for each cross-validation fold, Nfeats n. of selected features

ISSN: 1471-2105