EEG-Based Automatic Sleep Staging Using Ontology and Weighting Feature Analysis

Table 14

Performance comparison of various state-of-the-art methods in terms of accuracy (%).

Researchers	Five-state (%)	Four-state (%)	Three-state (%)	Two-state (%)
Sharma et al. [6]	91.70	92.10	93.90	98.30
Liang et al. [9]	83.60	—	—	—
Sharma et al. [10]	91.13	92.29	94.66	98.02
Zhu et al. [13]	88.90	89.30	92.60	97.90
Hsu et al. [17]	87.20	—	—	—
Berthomier et al. [23]	71.20	74.50	88.30	95.40
Our proposed method	90.57	92.41	94.85	97.21

Boldface indicates that the highest classification accuracy values. “—” indicates the missing cases.