Table 3 The hyper parameters of our model
From: Dependency-based long short term memory network for drug-drug interaction extraction
Parameter | Description | Value |
|---|---|---|
dw | Dimension of word embedding | 100 |
dp | Dimension of distance embedding | 10 |
num | The number of hidden units | 300 |
ρ | The ratio of dropout | 0.7 |
l 2 | The L2 regularization | 0.001 |
l a | The learning rate of Adam optimizer | 0.01 |
α | The ratio of under sampling | 0.5 |
β | The ratio of oversampling | 0.5 |
K | The times of oversampling | 6 |