Table 3.

Emotion prediction performance of machine learning algorithms in the frequency domain.

Class	Index	LR	NB	SVM	RF	KNN	CNN+LSTM	SNN	SBP-SNN
Fearful and Neutral	Acc. (%)	31.48	52.44	79.13	83.86	73.41	89.25	82.18	86.36
	F1-Sc.	0.31	0.61	0.53	0.79	0.64	0.83	0.82	0.92
Sad and Neutral	Acc. (%)	62.85	57.68	73.38	47.26	61.84	82.50	86.24	95.18
	F1-Sc.	0.43	0.43	0.80	0.29	0.57	0.79	0.85	0.95
Happy and Neutral	Acc. (%)	36.66	52.10	79.65	73.43	66.12	89.38	83.56	89.09
	F1-Sc.	0.28	0.28	0.70	0.75	0.64	0.79	0.85	0.80
Four classes	Acc. (%)	–	61.19	68.88	58.71	61.66	67.10	65.35	70.25
	F1-Sc.	–	0.48	0.63	0.47	0.55	0.65	0.70	0.73