Table 2.

Head and tail fusion model performance.

Fusion model					Precision (%)			Recall (%)			F-score (%)
Head entity fusion
	Linear regression models
		Jaa+TransR	83.37			84.06			83.71
		Sab+TransR	83.37			84.55			83.95
		Ja+Sa	83.85			84.55			84.19
		Ja+Sa+TransR	83.92			84.61			84.26
	Neural network			97.29			97.03			97.16
	Decision tree			97.47			97.18			97.32
Tail entity fusion
	Linear regression models
		BERTc+TransR	61.73			61.74			61.73
		Ja+BERT	95.76			95.83			95.79
		Ja+TransR	95.89			95.93			95.90
		Ja+BERT+TransR	95.92			95.94			95.93
	Neural network			98.43			98.72			98.57
	Decision tree			98.18			98.05			98.11

^aJa : Jaccard similarity.

^bSa: identical tail nodes in different sets: .

^cBERT: bidirectional encoder representations from transformer.