Table 1 Comparison results of EviDTI and baselines on the DrugBank dataset

From: Evidential deep learning-based drug-target interaction prediction

DrugBank

Accuracy (std)

Recall (std)

Precision (std)

MCC (std)

F1 score (std)

AUC (std)

AUPR (std)

RF

71.07 (1.34)

73.08 (0.95)

76.92 (1.42)

51.23 (1.54)

74.94 (0.92)

83.03 (0.65)

84.78 (0.67)

NB

52.90 (1.13)

87.99 (1.77)

51.71 (1.01)

42.34 (3.24)

65.13 (0.89)

60.76 (1.24)

59.37 (1.77)

SVM

78.02 (1.92)

75.04 (2.93)

73.86 (2.14)

53.87 (1.29)

74.39 (1.40)

82.02 (1.62)

82.85 (1.49)

DeepConv-DTI

79.50 (0.22)

80.45 (0.21)

79.30 (0.64)

57.91 (0.61)

79.87 (0.62)

86.69 (0.22)

86.28 (0.13)

GraphDTA

77.69 (0.11)

78.77 (1.13)

76.92 (1.05)

55.41 (0.63)

77.83 (1.18)

83.62 (0.25)

83.26 (0.18)

MolTrans

78.48 (0.17)

81.97 (0.23)

77.01 (0.35)

55.65 (0.62)

79.41 (0.73)

85.98 (0.64)

86.67 (0.12)

HyperAttention

80.85 (0.23)

88.16 (0.15)

77.09 (0.17)

62.32 (0.42)

82.26 (0.33)

89.94 (0.14)

88.95 (0.15)

TransformerCPI

79.01 (0.24)

76.26 (0.36)

80.47 (0.34)

58.09 (0.22)

78.31 (0.38)

84.62 (0.33)

84.23 (0.28)

GraphormerDTI

80.22 (1.07)

82.05 (1.23)

79.11 (0.91)

60.48 (2.16)

80.55 (1.02)

88.63 (1.02)

88.36 (0.99)

AIGO-DTI

81.88 (1.37)

82.28 (2.69)

81.26 (0.36)

63.80 (2.78)

81.74 (1.40)

88.62 (1.02)

88.44 (1.65)

DLM-DTI

82.13 (0.27)

82.54 (1.91)

81.68 (1.56)

66.32 (0.55)

81.08 (0.56)

90.48 (0.48)

90.20 (0.60)

EviDTI

82.02 (0.43)

82.28 (1.03)

81.90 (0.58)

64.29 (1.05)

82.09 (0.74)

88.87 (0.73)

88.45 (1.51)

  1. The best results are highlighted in bold, while the second-best results are underlined. Five independent replications of each method were performed (n = 5). Data are expressed as means (std).