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

Table 1

The main layers and their optional parameters in the deep learning approach.

Layer	Parameter
Convolution1D_1	Filter (8, 16, 32, and 64), kernel size (2, 5, 10, and 15), strides (1)
MaxPooling_1	Pool size (2), strides (1)
Convolution1D_2	Filter (8, 16, 32, and 64), kernel size (2, 5, 10, and 15), strides (1)
MaxPooling_2	Pool size (2), strides (1)
BLSTM_1	Memory cell (32, 64, 128, and 256)
BLSTM_2	Memory cell (32, 64, 128, and 256)
Dense_1	Output space (64)
Dropout_1	Rate (0.6)
Dense_2	Output space (25)
Dropout_2	Rate (0.6)
Softmax	Output space (1)