Skip to main content

Table 8 Standard performance metrics for Machine Learning models for the prediction of PCa aggressiveness

From: Development of a predictive model to distinguish prostate cancer from benign prostatic hyperplasia by integrating serum glycoproteomics and clinical variables

Algorithm

AUC

f1

Accuracy

Specificity

Sensitivity

Random forest

0.69 (0.57–0.81)

0.69 (0.57–0.81)

0.67 (0.54–0.79)

0.60 (0.47–0.73)

0.78 (0.67–0.89)

Logistic regression

0.50 (0.37–0.63)

0.77(0.66–0.88)

0.63 (0.50–0.75)

1

0

KNN

0.57 (0.44–0.70)

0.73 (0.61–0.84)

0.63 (0.5–0.75)

0.80 (0.69–0.91)

0.33 (0.21–0.46)

Decision tree

0.43 (0.30–0.56)

0.29 (0.17–0.41)

0.38 (0.25–0.50)

0.20 (0.09–0.31)

0.67 (0.54–0.79)