Table 3 Parameters for proposed deep stacked sparse autoencoder neural network.

From: Breath analysis based early gastric cancer classification from deep stacked sparse autoencoder neural network

Model number

Learning rate

Hidden size

Sparsity regularization

Sparsity proportion

1st AE

2nd AE

1st AE

2nd AE

1st AE

2nd AE

1st AE

2nd AE

1

0.001

0.001

100

20

4

4

0.05

0.05

2

0.001

0.001

100

40

4

4

0.05

0.05

3

0.001

0.001

100

60

4

4

0.05

0.05

4

0.001

0.001

100

80

4

4

0.05

0.05

5

0.001

0.001

100

100

4

4

0.05

0.05

Back to article page