Recognition of Emotions Using Multichannel EEG Data and DBN-GC-Based Ensemble Deep Learning Framework

Table 4

Average recognition performance comparison between the DBN-GC-based ensemble deep learning model and several reported studies.

	Arousal		Valence
	Accuracy	F1-score	Accuracy	F1-score
Deep neural network [29]	0.7313	—	0.7578	—
Convolutional neural network [29]	0.7336	—	0.8141	—
CNN/RNN [37]	0.7412	—	0.7206	—
DBN-SVM [37]	0.6420	—	0.5840
Wang and Shang [26]	0.5120	—	0.6090	—
CNS feature-based single modality [32]	0.6200	0.5830	0.5760	0.5630
DBN-GC-based ensemble deep learning model	0.7592	0.6931	0.7683	0.7015