Skip to main content

Table 5 The detailed parameters of the model

From: Sleep postures monitoring based on capacitively coupled electrodes and deep recurrent neural networks

Parameter

Value

Input

[ECG_channel1, ECG_channel2]

Output

Category of the sleep posture (Sp, Lp, Rp)

Layer number of biLSTM

3

biLSTM size

[200, 100, 50]

biLSTM state activation function

tanh

biLSTM gate activation function

Sigmoid

Output layer

Softmax

Loss function

Cross-entropy loss function

Optimizer

Adam

Number of training epochs

150