Table 1.

Results of event extraction on the test sets given gold entities

Task	Model	P	R	F (%)
CG	TEES-CNN (Björne and Salakoski, 2018)	66.55	50.77	57.60
	DeepEventMine (single)	69.54	54.24	60.94
	DeepEventMine (ensemble)	72.23	53.92	61.74
EPI	EventMine (Miwa et al., 2013b)	54.42	54.28	54.35
	TEES-CNN (Björne and Salakoski, 2018)	64.93	50.00	56.50
	DeepEventMine (single)	73.73	55.95	63.62
	DeepEventMine (ensemble)	78.34	56.39	65.57
GE11	EventMine (Miwa et al., 2012)	63.48	53.35	57.98
	BioMLN (Venugopal et al., 2014)	63.61	53.42	58.07
	TEES-CNN (Björne and Salakoski, 2018)	69.45	49.94	58.10
	DeepEventMine (single)	71.71	56.20	63.02
	DeepEventMine (ensemble)	76.28	55.06	63.96
GE13	TEES-CNN (Björne and Salakoski, 2018)	65.78	44.38	53.00
	BioMLN (Venugopal et al., 2014)	59.24	48.95	53.61
	DeepEventMine (single)	60.98	49.80	54.83
	DeepEventMine (ensemble)	67.08	49.14	56.72
ID	TEES-CNN (Björne and Salakoski, 2018)	66.48	50.66	57.50
	EventMine (Miwa et al., 2013b)	61.33	58.96	60.12
	DeepEventMine (single)	63.56	57.30	60.27
	DeepEventMine (ensemble)	68.51	55.99	61.62
PC	EventMine (Miwa and Ananiadou, 2013)	53.48	52.23	52.84
	TEES-CNN (Björne and Salakoski, 2018)	62.16	50.34	55.62
	DeepEventMine (single)	64.12	49.19	55.67
	DeepEventMine (ensemble)	68.13	50.07	57.72
MLEE	MultiRep-CNN (Wang et al., 2017)	60.56	56.23	58.31
	PMCNN (Li et al., 2020)	67.23	53.61	59.65
	BLSTM (He et al., 2019)	—	—	59.61
	DeepEventMine (single)	67.39	56.35	61.38
	DeepEventMine (ensemble)	69.91	55.49	61.87

Notes: The highest scores are shown in bold.