EEG-Based Emotion Recognition Using Deep Learning Network with Principal Component Based Covariate Shift Adaptation

Table 1

DLN parameter settings.

Parameters	Value
Maximum iterations: SAE learning	400
Maximum iterations: softmax learning	100
Maximum iterations: fine-tuning	200
Hidden layer size	100, 50
Sparsity parameter (	0.10
Weight of sparsity penalty (β)	3.0
Weight decay parameter (): SAE learning
Weight decay parameter (): softmax learning
Weight decay parameter (): fine-tuning