Evolutionary Multilabel Feature Selection Using Promising Feature Subset Generation

Table 2

Comparison results in terms of multilabel accuracy.

Method	Birds	Enron	Llog	Mediamill	Medical
Proposed	0.527 ± 0.044	0.383 ± 0.012	0.249 ± 0.014	0.362 ± 0.004	0.427 ± 0.030
GA	0.491 ± 0.055	0.284 ± 0.022	0.209 ± 0.018	0.336 ± 0.031	0.303 ± 0.058
NSGA	0.480 ± 0.040	0.282 ± 0.022	0.208 ± 0.014	0.347 ± 0.013	0.297 ± 0.018
MPSOFS	0.453 ± 0.031	0.206 ± 0.012	0.042 ± 0.002	0.163 ± 0.007	0.286 ± 0.033
Method	TMC2007	Business	Education	Entertainment	Health
Proposed	0.441 ± 0.004	0.672 ± 0.011	0.318 ± 0.009	0.396 ± 0.004	0.537 ± 0.011
GA	0.435 ± 0.010	0.657 ± 0.021	0.316 ± 0.010	0.361 ± 0.011	0.499 ± 0.011
NSGA	0.434 ± 0.005	0.662 ± 0.013	0.318 ± 0.010	0.362 ± 0.010	0.495 ± 0.009
MPSOFS	0.420 ± 0.005	0.634 ± 0.008	0.283 ± 0.005	0.365 ± 0.010	0.496 ± 0.013
Method	Reference	Science	Social	Society	Avg. rank
Proposed	0.436 ± 0.009	0.288 ± 0.010	0.546 ± 0.008	0.371 ± 0.008	1.00
GA	0.422 ± 0.012	0.231 ± 0.009	0.517 ± 0.012	0.258 ± 0.011	2.79
NSGA	0.429 ± 0.009	0.237 ± 0.010	0.526 ± 0.009	0.267 ± 0.010	2.64
MPSOFS	0.414 ± 0.017	0.234 ± 0.013	0.527 ± 0.012	0.239 ± 0.007	3.57