Table 12 Accuracy improvement in eight combinations when NSGA-II is used for EEG channel and feature selection.

From: Selecting EEG channels and features using multi-objective optimization for accurate MCI detection: validation using leave-one-subject-out strategy

Combination

Classification Accuracy based on:

All channels (No_ch)

Subset of channels (No_ch)

Subset of features (No_feat, No_ch)

VMD-KFD-RF

74.24 (19)

85.49 (5)

90.83 (14,10)

VMD-TeEng-SVM

74.24 (19)

91.56 (5)

95.28 (8,7)

VMD-TeEng-DA

79.58 (19)

91.94 (5)

95.63 (15,10)

VMD-SuEn-KNN

62.36 (19)

86.04 (4)

91.53 (10,6)

DWT-ThEn-RF

68.54 (19)

87.64 (3)

90.63 (9,6)

DWT-TeEng-SVM

80.14 (19)

92.64 (5)

95.83 (9,8)

DWT-ThEn-DA

75.42 (19)

94.65 (8)

95.63 (13,8)

DWT-TeEng-KNN

58.54 (19)

85.35 (2)

93.13 (10,6)

  1. No_ch: number of selected channels, and No_feat: number of selected features.
Back to article page