A Deep Learning Approach for Predicting Antigenic Variation of Influenza A H3N2

Table 5

Comparison between the agreements obtained by our deep learning approach and the Antigen-Bridges method with three residue sets [21] for the strains in the forthcoming year and in the next two years.

Approaches (amino acid number)	Next year (%)	Next two years (%)
Antigen-Bridges (39-residue set)	83.78	75.10
Antigen-Bridges (44-residue set)	79.75	72.48
Antigen-Bridges (25-residue set)	80.51	71.51
Our deep learning approach	99.20	96.46