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

Table 6 SNFS using CDD versus several approaches using SVM on leukemia and colon cancer datasets

Reference	CS	Dataset	FS	Time	FN	Acc.	AUC
Our approach	1	Leukemia	SNFS with CDD	71.22 (59.04-83.38)\(^1\)	8-10	88.24-97.06\(^2\)	93.93-100\(^2\)
Our approach	1	Colon cancer	SNFS with CDD	37.23 (32.33-83)\(^1\)	11-12	63.16-89.47\(^2\)	82.14-96.43\(^2\)
Ozyer et al. [5]	2	Leukemia	SNFS	NM	9	97.06	NM
Ozyer et al. [5]	2	Colon cancer	SNFS	NM	9	100	NM
Ding et al. [56]	3	Leukemia	OELM-RFE	NM	4	84.38	NM
Khaire and Dhanalakshmi [57]	NM	Leukemia	IWOA	NM	NM	81	92
Khaire and Dhanalakshmi [57]	NM	Colon cancer	IWOA	NM	NM	83	84
Kundu et al. [58]	4	Leukemia	AWOA	41.53	30	100	100
Kundu et al. [58]	4	Colon cancer	AWOA	10.37	21	100	Approx. 100
Vatankhah and Momenzadeh [59]	5	Leukemia	Self-regularized LASSO	14.97	1	NM	100

CDD: corrected degree of domesticity; CS: computational resources; FS: feature selection method; time: execution time in seconds; FN: number of selected features; Acc.: accuracy; AUC: area under the ROC curve; NM: Not mentioned; Approx.: approximately
CS: 1: Intel(R) Core(TM) i5-8400 CPU @ 2.80 GHz 8.00 GB of RAM; 2: Intel(R) Core(TM) i5 CPU @ 3.10 GHz with 4.00 GB of RAM; 3: I7-7700HQ CPU @ 2.8 GHz with 8.00 GB of RAM; 4: Intel(R) Core(TM) i5 with 8.00 GB of RAM; 5: Intel(R) Core(TM) i7-10750 H CPU @ 2.60 GHz with 16.00 GB of RAM
\(^1\)Median (minimum-maximum) values for execution time of the possible combinations in SNFS
\(^2\)Minimum-maximum values of the performance measure obtained from the possible combinations in SNFS

ISSN: 1471-2105