Table 2 Evaluation of the classification performance of PMFs for differentiating healthy individuals from hepato-pancreato-biliary cancer patients using a support vector machine model.

From: Serum peptide biomarkers by MALDI-TOF MS coupled with machine learning for diagnosis and classification of hepato-pancreato-biliary cancers

Group

Support vector machine

ACC (%)

Precision (%)

Recall (%)

F1-Score

SPC (%)

MCC

ROC

Training set (n = 198)

Healthy (n = 29) vs. All (n = 169)

98.74

99.70

98.82

0.99

98.28

0.95

0.999

Healthy vs. CCA (n = 97)

100

100

100

1

100

1

1

Healthy vs. GBC (n = 8)

100

100

100

1

100

1

0.998

Healthy vs. HCC (n = 45)

100

100

100

1

100

1

1

Healthy vs. PDAC (n = 19)

99.48

98.70

100

0.99

99.14

0.99

0.999

Testing set (n = 99)

Healthy (n = 21) vs. All (n = 78)

98.55

99.35

97.76

0.99

99.35

0.97

0.999

Healthy vs. CCA (n = 41)

99.60

100

99.39

1

100

0.99

1

Healthy vs. GBC (n = 8)

97.41

93.94

96.88

0.95

97.62

0.94

0.974

Healthy vs. HCC (n = 20)

100

100

100

1

100

1

1

Healthy vs. PDAC (n = 9)

97.50

97.14

94.44

0.96

98.81

0.94

0.985