Attention-based deep residual learning network for entity relation extraction in Chinese EMRs

Table 6 Comparison of F₁-score for each model on every fine-grained relation category

Category	SVM	CNN-Max	LSTM-Max	BLSTM-Attention	ResNet-BLSTM	ResNet-Max	ResNet-PAtt
TrID	20.06	29.68	36.78	40.38	42.67	35.42	44.62
TrWD	19.35	28.34	25.35	35.21	33.21	30.43	40.73
TrCD	28.52	28.02	39.41	46.32	48.57	42.6	46.50
TrAD	63.21	43.52	58.31	71.65	68.33	64.54	70.39
TrNAD	12.36	22.46	18.24	36.69	37.26	35.42	46.03
TrIS	57.24	48.52	49.31	54.37	52.44	52.31	56.53
TrWS	36.41	49.51	37.53	46.21	48.18	42.43	49.23
TrCS	39.04	39.53	41.52	39.46	40.93	39.5	40.39
TrAS	60.26	58.33	62.34	66.83	72.36	61.36	69.67
TrNAS	13.54	14.39	14.56	28.67	30.31	24.67	36.13
TeRD	62.35	60.27	62.24	71.36	74.22	69.96	73.64
TeCD	12.34	16.52	18.36	37.23	32.88	31.48	40.42
TeRS	82.53	71.26	75.34	80.44	81.63	78.45	83.61
TeBS	48.42	46.34	47.21	58.64	57.94	51.20	57.50
DCS	64.28	65.31	65.67	74.24	73.55	70.69	75.68

ISSN: 1472-6947