Skip to main content

Table 1 The CNN parameters

From: An interpretable DIC risk prediction model based on convolutional neural networks with time series data

Layer

Filters

Size

Stride

Output

Conv1(convolution  + BatchNorm + ReLu

16

(3,1)

(2,1)

16 * 476 * 136

Conv2(convolution  + BatchNorm + ReLu

32

(3,1)

(2,1)

32 * 237 * 136

Conv3(convolution  + BatchNorm + ReLu

64

(3,1)

(2,1)

64 * 118 * 136

Maxpool

 

(118,1)

 

64 * 1 * 136

Conv4(convolution  + ReLu)

1

(1,1)

(1,1)

1 * 1 * 136

Fully-connected layer

–

–

–

136

Output layer

–

–

–

2