Table 5 Comparison of the three-classification between STFFDA and other models with 10-fold cross-validation.

From: EEG detection and recognition model for epilepsy based on dual attention mechanism

Methods

Accuracy(%)

Precision(%)

Recall(%)

F1-score(%)

MCC(%)

DNN

56.35

56.20

56.38

55.96

34.73

CNN

60.80

61.46

60.80

60.44

41.61

CNN-RNN

57.17

57.15

57.20

56.89

35.95

1D Inception-v1

65.95

66.05

65.94

65.66

49.14

Decision Tree

71.35

71.38

71.36

71.36

57.04

KNN

91.44

91.75

91.45

91.46

87.29

Bayes

45.54

46.38

45.61

43.06

19.85

STFFDA

92.42

92.47

92.42

92.42

88.66

  1. The table summarizes the average results of the ten-fold cross-validation for the three-class classification task on the CHB-MIT dataset. STFFDA leads again with an accuracy of 92.42%, demonstrating its stable performance across multiple folds.
Back to article page