From: Medication adherence prediction through temporal modelling in cardiovascular disease management
Models | Hyperparameters |
---|---|
LSTM | Layers: 1 LSTM and 1 Dense |
 | Units: 4 (LSTM) and 2 (Dense) |
 | Batch size: 1024 |
 | L2: 9.261e−3 |
 | Loss: categorical cross-entropy |
 | Epochs: 100 |
Simple RNN | Layers: 1 Simple RNN and 1 Dense |
 | Units: 8 (Simple RNN) and 2 (Dense) |
 | Batch size: 4096 |
 | L2: 1.202e−2 |
 | Loss: categorical cross-entropy |
 | Epochs: 100 |
MLP | Layers: 3 Dense and 2 Dropout |
 | Units: 64, 64 and 2 |
 | Batch size: 256 |
 | Dropout rate: Layer 1 2.152e−1 |
 | Layer 2 1.758e−1 |
 | Loss: categorical cross-entropy |
 | Epochs: 50 |