Automatic text classification of actionable radiology reports of tinnitus patients using bidirectional encoder representations from transformer (BERT) and in-domain pre-training (IDPT)

Table 8 Comparison of BERT finetune with different max sequence lengths

Model	Max sequence length	Accuracy	Precision	Recall	AUC	F1-score
BERT fine-tune	128	0.741	0.738	0.741	0.866	0.736
	256	0.71	0.707	0.71	0.843	0.708
	328	0.616	0.627	0.616	0.797	0.601
	468	0.551	0.557	0.551	0.759	0.546
	512	0.760	0.761	0.759	0.868	0.760

ISSN: 1472-6947