Table 4.

Characteristics of several multi-label data sets compared with ours.

Name	Complexity	N	|L|	d	d/N	LCavg	PUNIQ	Pmax
StimModAbsa	4,449,705	247	5	3603	14.59	1.15	0.036	0.008
CogParaAllb	46,451,808	247	48	3919	15.87	1.13	0.336	0.004
Medical	63,770,490	978	45	1449	1.48	1.25	0.096	0.158
Slashdot	89,777,116	3782	22	1079	0.29	1.18	0.041	0.139
Enron	90,296,206	1702	53	1001	0.59	3.38	0.442	0.096

Values taken from Read et al. (2011); see there for details and sources. For notation, see section 2.1.2 and 2.2. Included are the values for the least and most complex data sets included in this paper.

^aAbstract alone corpus; stimulus modality labels.

^bAbstract, title, and keyword corpus; cognitive paradigm class labels.