Table 2 Number of training parameters of the GCN model according to the Chebyshev polynomial order of each layer.

From: An interpretable model based on graph learning for diagnosis of Parkinson’s disease with voice-related EEG

Layers

Shape of weight tensor

Shape of bias

Number of parameters

Convolution layer1

[K1, 64, 64]

[64]

4096×K1 + 64

Pool1

[1×2]

[1×2]

4

Convolution layer2

[K2, 64, 32]

[32]

2048×K2 + 32

Pool2

[2×1]

[2×1]

4

Fully Connected layer

[32, 32]

[32]

1056

  1. K1 and K2 are the coefficients of the Chebyshev polynomial expansion.
Back to article page