Open Access
Table 3
Neural network hyper-parameters.
| Hyper-parameter | Value | |
|---|---|---|
| Architecture | No. of hidden layers | 3 |
| Neurons per hidd. layer | 64 | |
| Hidd. Neuron gate | ReLU | |
| Hidd. layer type | Dense | |
| Output layer gate | Sigmoid | |
| Training | Data split | 70-15-15% |
| Seed | 42 | |
| Shuffle | True | |
| Learning rate | 10−4 | |
| Batch size | 64 | |
| Optimizer | Adam | |
| Loss function | Negative log-loss | |
| Overfitting | Dropout rate | 0.2 |
| EarlyStopping(monitor) | val data loss | |
| EarlyStopping(patience) | 10 |
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.