Skip to main content

Table 4 Performance of various down-sampling methods on independent test dataset

From: GBDT_KgluSite: An improved computational prediction model for lysine glutarylation sites based on feature fusion and GBDT classifier

Model

Samples (N/P)

Acc (%)

Sen (%)

Pre (%)

F1 (%)

MCC

AUC (%)

ClusterCentroids

530/530

84.75

81.77

87.57

84.57

69.68

92.55

RUS

530/530

84.18

81.22

86.98

84.00

68.55

92.18

OneSideSelection

2863/530

82.20

80.66

83.91

82.25

64.47

90.21

NearMiss-1

530/530

83.62

80.11

86.83

83.33

67.48

93.11

NearMiss-2

530/530

85.31

83.43

87.28

85.31

70.71

92.00

NearMiss-3

530/530

90.11

95.06

85.08

89.79

80.73

96.75

  1. To facilitate understanding, the highest value in each column is shown in bold. where the N and P in the Samples column brackets means negative and positive, respectively