Skip to main content

Table 7 Network parameters of the Bi-LSTM

From: Attention-based multi-scale features fusion for unobtrusive atrial fibrillation detection using ballistocardiogram signal

Layer

Size

Description

Input

125

Input signal length

Bi-LSTM

50

First Bi-LSTM layer

Dropout

50

Random dropout to avoid overfitting

Bi-LSTM

50

Second Bi-LSTM layer

FC

2

Map the features

Softmax

2

Softmax function

Output

2

Output the classification results