Table 10.

Comparative analysis of the methods and outcomes of the proposed work with other sleep studies.

Study	Year	Study Subject	Dataset (Year)/Signal	Class	Algorithm	Accuracy %
Tzimourta et al. [40]	2018	100 subjects	ISRUC-Sleep dataset (2009–2013)/EEG	Five-class {W, N1, N2, N3, and REM}	Random Forest	75.29
Kalbkhani et al. [41]	2018	100 subjects	ISRUC-Sleep dataset (2009–2013)/EEG	Five-class {W, N1, N2, N3, and REM}	SVM	82.33
Tripathi et al. [42]	2020	25 subjects	Cyclic Alternating Pattern (CAP) (2001)/EEG	Six-class {W, S1, S2, S3, S4, and REM}	Hybrid Classifier	71.68
Widasari et al. [43]	2020	51 subjects	Cyclic Alternating Pattern (CAP) (2001)/EEG	Four-class {W, Light sleep (S1 + S2), Deep sleep (S3 + S4), and REM}	Ensemble of bagged tree (EBT)	86.26
Wang et al. [44]	2020	157 subjects	Sleep-EDF Expanded (Sleep-EDFX) (2000)/EEG and EOG	Five-class {W, N1, N2, N3, and REM}	Ensembles of EEGNet-BiLSTM	82
Sharma et al. [45]	2021	80 subjects	Cyclic Alternating Pattern (CAP) (2001)/EEG	Six-class {W, S1, S2, S3, S4, and REM}	Ensemble of Bagged Tree (EBT)	85.3
Proposed work	2022	157 subjects	HMC-Haaglanden Medisch Centrum (2021)/EEG	Five-class {W, N1, N2, N3, and REM}	C5.0, Neural Network, and CHAID	C5.0 (91%), Neural Network (92%), and CHAID (84%)