Skip to main content

Table 6 The efficacy of ANN, logistic regression, kNN and SVM in testing sets

From: Application of artificial neural network model in diagnosis of Alzheimer’s disease

Assessment index

ANN (95% CI)

Logistic regression (95% CI)

kNN (95% CI)

SVM (95% CI)

Sensitivity (%)

87.28 (85.12–88.34)

71.36 (69.18–72.93)

75.42 (73.24–76.17)

84.23(82.78–85.63)

Specificity (%)

94.74 (92.61–95.75)

90.25 (87.82–91.85)

89.39 (87.65–90.49)

90.45(88.65–91.73)

Accuracy (%)

92.13 (89.48–92.57)

83.92 (80.31–85.11)

84.91 (83.01–86.41)

86.04(84.93–88.10)

PPV (%)

88.64 (86.83–90.06)

73.11 (71.09–74.86)

83.02 (82.17–85.23)

82.36 (81.25–83.94)

NPV (%)

91.28 (90.41–93.10)

86.37 (85.19–88.26)

86.41 (84.85–88.15)

88.77(87.21–90.33)

AUC

0.897 (0.877–0.915)

0.804 (0.781–0.818)

0.832 (0.817–0.849)

0.864 (0.853–0.876)