Multi-Label Feature Selection with Conditional Mutual Information

Table 2

Hamming Loss of results after applying multi-label feature selection algorithms on ten datasets.

Algorithms	CRMIL	SCLS	D2F	FIMF	PMU	AMI	NMDG	FSSL	MFS-MCDM
Corel5k	0.0104	0.0144	0.0235	0.0226	0.0227	0.0136	0.0141	0.0138	0.0194
Delicious	0.0170	0.0229	0.0336	0.0398	0.0374	0.0195	0.0217	0.0203	0.0231
Flags	0.3165	0.4022	0.3605	0.4132	0.3561	0.3496	0.3855	0.3520	0.3935
Medical	0.0211	0.0279	0.0386	0.0238	0.0319	0.0274	0.0253	0.0251	0.0283
Scene	0.2413	0.3009	0.3016	0.3363	0.3083	0.3126	0.2783	0.2679	0.2804
Enron	0.0723	0.0973	0.1027	0.0989	0.1031	0.0974	0.0847	0.0811	0.0873
GenBase	0.0052	0.0079	0.0062	0.0103	0.0091	0.0098	0.0074	0.0077	0.0101
Social	0.0424	0.0512	0.0563	0.0534	0.0712	0.0491	0.0472	0.0469	0.0507
Yeast	0.2319	0.2512	0.2579	0.2603	0.2591	0.2487	0.2449	0.2433	0.2496
Emotions	0.2613	0.2817	0.2833	0.3074	0.3096	0.2913	0.2804	0.2716	0.2775
Average	0.1219	0.1458	0.1464	0.1566	0.1651	0.1419	0.1390	0.1330	0.1420