Table 1. Per-class performance of the conclusiveness classification for Scientific Bidirectional Encoder Representations from Transformers (SciBERT) and Longformer models^a.

Model	Precision	Recall	F1-score	AUCROCb score
Conclusive
SciBERT	0.43	0.58	0.50	0.91
Longformer	0.43	0.74	0.55	0.87
Inconclusive
SciBERT	0.42	0.35	0.38	0.67
Longformer	0.46	0.31	0.37	0.67
Unclear
SciBERT	0.75	0.77	0.76	0.75
Longformer	0.75	0.78	0.76	0.72

^aPrecision is the proportion of true positives relative to the total of true positives and false positives. Recall is the proportion of true positives relative to the total of true positives and false negatives. F₁-score is the harmonic mean of the precision and recall. Area under the curve of the receiver operating characteristic is the proportion of area below the receiver operating characteristic curve, which is the plot of the true positive rate against the false positive rate. The scores were obtained using the scikit-learn functions.

^bAUCROC: area under the curve of the receiver operating characteristic.