Table 3.

Comparison of different metadata encoders (private dataset)

Methods	Acc1	Acc2	Spe	Sen	F1	AUC
Word2vec	0.313	0.750	0.967	0.746	0.718	0.944
One-hot Encoder	0.401	0.763	0.962	0.783	0.745	0.947
Soft Label Encoder	0.473	0.777	0.973	0.804	0.788	0.964

The bold values represent the best results in the metrics. This can make more readable and make it easier to compare the results of different methods. This also highlights the effectiveness of the proposed method