Table 2 Performance of 5-fold cross-validation for the Bayesian logistic regression and FLR models

From: Establishment of a machine learning model for predicting splenic hilar lymph node metastasis

 

ROC AUC

PRAUC

Sensitivity

Specificity

Accuracy

Precision

F1 score

Bayes-SHLNM

0.83

(0.74–0.91)

0.35

(0.14–0.56)

0.73

(0.54–0.92)

0.74

(0.69–0.80)

0.74

(0.70–0.79)

0.20

(0.14–0.26)

0.31

(0.23–0.40)

Basic

0.77

(0.66–0.89)

0.26

(0.09–0.43)

0.50

(0.19–0.82)

0.83

(0.76–0.90)

0.80

(0.74–0.86)

0.20

(0.90–0.32)

0.28

(0.12–0.45)

Student-T

0.78

(0.67–0.89)

0.26

(0.09–0.43)

0.50

(0.19–0.82)

0.82

(0.75–0.89)

0.79

(0.74–0.85)

0.19

(0.09–0.30)

0.27

(0.12–0.43)

Bayesian LASSO

0.78

(0.67–0.89)

0.26

(0.10–0.43)

0.50

(0.19–0.82)

0.82

(0.75–0.89)

0.79

(0.74–0.85)

0.19

(0.09–0.30)

0.27

(0.12–0.42)

FLR

0.83

(0.76–0.90)

0.37

(0.13–0.61)

0.69

(0.61–0.77)

0.77

(0.70–0.85)

0.77

(0.70–0.83)

0.22

(0.17–0.27)

0.33

(0.28–0.38)

  1. ROC area under the receiver operating characteristic curve, PRAUC area under the precision-recall curve, Bayes-SHLNM Bayes-splenic hilum lymph node metastasis, LR logistic regression, LASSO least absolute shrinkage and selection operator; Bayesian logistic regression with horseshoe prior. Performance metrics of various logistic regression models in predicting SHLN metastasis in patients with UGC. The results are based on 5-fold cross-validation and represent the mean values with 95% CI. The models compared included the frequentist logistic regression (FLR) model, basic Bayesian model, Bayesian Student-T model, Bayesian LASSO model, and Bayesian logistic regression model with a horseshoe prior (Bayes-SHLNM model). The metrics included ROC AUC, PR AUC, sensitivity, specificity, accuracy, precision, and F1 score.
Back to article page