Automated classification of clinical trial eligibility criteria text based on ensemble learning and metric learning

Table 3 Performance comparison of all single models with or without metric learning using macro F1 score (margin parameter m = 0.1)

Model	Without metric learning	With metric learning	Increase rate (%)
BERT	0.7880	0.7973	1.18
XLNet	0.7983	0.8030	0.59
RoBERTa	0.7951	0.7992	0.52
ERNIE	0.7865	0.7952	1.11
ELECTRA	0.7758	0.7862	1.34

ISSN: 1472-6947