EEG Feature Extraction and Data Augmentation in Emotion Recognition

Table 8

Comparison of proposed work with similar work on DNN classifier.

Model	Features	NO. augmented data	Classification type	Mean accuracy	STD	Arousal		Valence
						M.Acc	STD	M.Acc	STD
Proposed model	344 extracted features	0	Binary	—	—	65.4%	% 3.4	64.3%	% 4.3
Proposed model	344 extracted features	2 real data	Binary	—	—	65.2%	% 4.2	62.3%	% 4.7
Proposed model	344 extracted features	Real data + 5000	Binary	—	—	71.9%	%4.8	67.4%	%5.2
[1]	DE	0	Categorical	44.9%	4.0%	—	—	—	—
[1]	DE	5000	Categorical	46.9%	4.8%	—	—	—	—