Selecting Optimal Feature Set in High-Dimensional Data by Swarm Search

Table 6

Comparative results in time consumption (in seconds) for all the classification algorithms with Cfs feature selection and swarm search feature selection and without any feature selection.

Algorithm	Lung Cancer (56)	Heart Disease (75)	Libra Movement (91)	Hill Valley (101)	CNAE (875)
PN	3.07	2.35	18.08	58.73	5066.72
DT	0.02	0.08	0.32	0.16	3.39
NB	≈0	0.01	0.05	0.08	0.26
Cfs-PN	0.18	≈0	3.79	0.3	10.51
Cfs-DT	≈0	0.01	0.04	≈0	0.08
Cfs-NB	≈0	1.26	0.02	≈0	≈0
PSO-PN	853.23123	382.3217	1714.34495	921.28525	70538.3472
PSO-DT	177.3407	11.9377	232.6246	491.7232	2886.9232
PSO-NB	22.9799	3.92775	82.7123	34.344	2358.2769
BAT-PN	1532.321	347.0041	686.557	524.367	26365.1231
BAT-DT	179.1062	16.5652	240.737	448.7332	2341.6668
BAT-NB	25.8876	2.436514	28.6922	31.001802	1441.7241
WSA-PN	1521.8085	1305.6519	3005.4028	1159.367	26365.1231
WSA-DT	1521.8085	49.2341	106.6606	176.6559	82688.1231
WSA-NB	42.2628	23.3551	404.6113	24.7167	236365.123