BMC Medical Informatics and Decision Making

Table 2 MISTIC performance metrics fine-tuned on 550 and 1100 training samples

From: MISTIC: a novel approach for metastasis classification in Italian electronic health records using transformers

	FT (550 samples)				FT (1100 samples)
Data source	P	R	F1	Acc	P	R	F1	Acc
Clinical diary	0.923	0.960	0.941	0.910	0.923	0.960	0.941	0.910
Medical history	0.789	0.984	0.876	0.830	0.782	1.000	0.878	0.830
Radiodiagnostic	0.841	1.000	0.914	0.860	0.830	0.986	0.901	0.840
Overall	0.851	0.981	0.912	0.867	0.844	0.981	0.907	0.860

FT fine-tuned model
P precision, R recall, F1 F1-score, Acc accuracy

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com