Open Access

Table 2.

CNN architecture and hyperparameter settings.

Input Layer Size Kernel_number Stride Output Parameters
2000*1 Dense(128) 2000*128 256
2000*50 Conv1D 1*1 128 1 2000*128 6528
2000*128 MaxPooling1D 2*1 1 1000*128 0
1000*128 Droupt 20% 1000*128 0
1000*128 Conv1D 8*1 64 1 993*64 65 600
993*64 MaxPooling1D 2*1 1 496*64 0
496*64 Droupt 20% 496*64 0
496*64 Conv1D 8*1 32 1 489*32 16 416
489*32 Droupt 20% 489*32 0
489*32 Conv1D 8*1 16 1 482*16 4112
482*16 Flatten 7712 0
7712 Dense(100) 100 771 300
100 Dense(7) 7 707

Loss Mean squared error
Optimizer (Adam) Lr = 0.0003
Epoch 100
Batch size 16

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.