The Comparative Experimental Study of Multilabel Classification for Diagnosis Assistant Based on Chinese Obstetric EMRs

Table 1

Results with |L₂| = 233, K = 120, and T = 100.

Method	Feature	HL↓	C↓	OE↓	RL↓	AP↑
RAkEL	LDA	0.0085 ± 0.0002	124.5190 ± 2.9857	0.3479 ± 0.0192	0.2874 ± 0.0092	0.5727 ± 0.0090
	Word vector	0.0078 ± 0.0002	127.1671 ± 2.4166	0.2902 ± 0.0173	0.2984 ± 0.0104	0.5906 ± 0.0109
MLkNN	LDA	0.0078 ± 0.000	15.5416 ± 0.7277	0.2425 ± 0.0127	0.0292 ± 0.0009	0.6571 ± 0.0087
	Word vector	0.0067 ± 0.0002	13.6120 ± 0.6596	0.2015 ± 0.0101	0.0240 ± 0.0007	0.7272 ± 0.0081
CC	LDA	0.0093 ± 0.0002	109.6586 ± 2.9200	0.4908 ± 0.0150	0.2430 ± 0.0078	0.5073 ± 0.0097
	Word vector	0.0088 ± 0.0001	90.2732 ± 2.8796	0.4427 ± 0.0109	0.1960 ± 0.0070	0.5408 ± 0.0074
BP-MLL	LDA	0.0341 ± 0.0058	14.6960 ± 1.0139	0.2426 ± 0.0136	0.0276 ± 0.0020	0.6264 ± 0.0114
	Word vector	0.0244 ± 0.0012	12.7561 ± 0.7484	0.2431 ± 0.0136	0.0225 ± 0.0009	0.6588 ± 0.0091