Table 8. Related works for anxiety detection.

Ref.	Preprocessing	Feature extraction techniques	Class label	Classification algorithm	Accuracy
Baghdadi et al. (2021)	Pass-band filter, automatic artifact removal, Hamming-windowed Welch, blind source separation	Hjorth parameters, fractal dimension, band power, Hilbert-Huang spectrum, discrete wavelet transform, and quantitative statistical EEG features	Two-level anxiety	Stacked Sparse AutoEncoder	83.50%
Chatterjee, Gavas & Saha (2023)		Spatio-temporal transition-based features	Both two-level and four-level anxiety	KNN	83.8%
Syakiylla Sayed Daud, Sudirman & Wee Shing (2023)	Bandpass filter, ICA, automatic artifact removal	Time-frequency domains sampling technique and safe-level SMOTE	Four-level anxiety decision tree	SVM	86.0%
					79.2%
Our approach	Band pass, channel selection, normalization	DWT and safe-level SMOTE	Two-level anxiety	RF	87.5%