Cooperative co-evolution for feature selection in Big Data with random feature grouping

Journal of Big Data

Table 10 Summary of results for the Musk Dataset

Classifier	Accuracy (%)	Sensitivity (%)	Specificity (%)	No. of features
NB	84.04	85.88	73.94	166
NB + CCEAFS	85.13	91.11	52.31	25
NB + CCFSRFG-1	89.71	94.63	62.73	20
NB + CCFSRFG-2	82.90	86.19	64.90	56
SVM	94.82	99.05	71.58	166
SVM + CCEAFS	91.71	98.24	55.85	22
SVM + CCFSRFG-1	92.79	98.98	58.80	14
SVM + CCFSRFG-2	92.82	98.84	59.78	63
k-NN	95.76	97.12	88.30	166
k-NN + CCEAFS	94.83	97.12	82.30	18
k-NN + CCFSRFG-1	96.59	98.15	88.00	9
k-NN + CCFSRFG-2	95.74	97.10	88.30	61
J48	96.85	98.15	89.68	166
J48 + CCEAFS	95.57	97.94	82.60	14
J48 + CCFSRFG-1	97.09	98.80	87.71	9
J48 + CCFSRFG-2	96.30	98.14	86.23	65
RF	97.94	99.87	87.32	166
RF + CCEAFS	97.06	99.44	83.97	16
RF + CCFSRFG-1	96.86	98.98	85.25	7
RF + CCFSRFG-2	97.56	99.71	85.74	60
LR	95.29	97.94	80.83	166
LR + CCEAFS	91.10	97.42	54.77	23
LR + CCFSRFG-1	92.10	98.30	58.11	10
LR + CCFSRFG-2	93.10	97.58	68.53	62

Italic signifies the improvements of the proposed approach over existing techniques