Emotion Recognition of EEG Signals Based on the Ensemble Learning Method: AdaBoost

Table 5

Comparison of the proposed method with some existing methods.

Reference	Feature domain	Class	Classifier	Accuracy
				Valence (%)	Arousal (%)	Dominance (%)	Liking (%)
Hwang et al. [14]	Frequency	Binary	SVM	64.90	64.90	—	66.80
Yoon and Chung [17]	Frequency	Binary	Bayesian	70.90	70.10	—	—
You and Liu [19]	Frequency + time-frequency	Binary	SAE + LSTM	81.10	74.38	—	—
Zhan et al. [22]	Frequency	Binary	CNN	82.95	84.07	—	—
Aggarwal et al. [24]	Time	Binary	GBM	77.11	60.25	—	—
Paruiet al. [23]	Nonlinear	Binary	XGBoost + LightGBM + random forest	77.19	79.06	—	—
Our proposed method	Time + time − frequency + nonlinear	Binary	AdaBoost	85.57	88.36	88.70	83.23