Table 8 Result of e_inc_tbhiv_100k variable performances.

From: A comparative study on TB incidence and HIVTB coinfection using machine learning models on WHO global TB dataset

Model

Accuracy (%)

Precision (%)

Recall (%)

F1 Score (%)

ROC AUC Score (%) u

GB

98.58

98.32

98.73

98.53

98.58

CB

98.58

98.94

98.1

98.52

98.56

XGB

98.27

98.31

98.1

98.2

98.27

ET

98.27

98.51

97.89

98.2

98.26

RF

98.07

98.72

97.26

97.98

98.04

AB

97.87

97.48

98.1

97.79

97.87

BC

97.76

98.29

97.05

97.66

97.74

DT

96.85

97.84

95.57

96.69

96.8

KNN

61.59

60.67

57.59

59.09

61.44

LR

51.83

0.00

0.00

0.00

50.00

SVM

50.51

49.18

82.7

61.68

51.64

SGDC

48.17

48.17

100.0

65.02

50.0

GNB

48.07

48.05

95.99

64.04

49.76

  1. Table shows e_inc_tbhiv_100k variable performances using the WHO TB burden dataset published in 2023 to predict TB Incidence and HIV-TB co-infection.
  2. GB: Gradient Boosting, CB: CatBoost, XGB: XGB, ET: Extra Trees, RF: Random Forest, AB: AdaBoost, BC: Bagging Classifier, DT: Decision Tree, KNN: K-Nearest Neighbors, LR: Logistic Regression, SVM: Support Vector Machine, SGDC: Stochastic Gradient Descent Classifier, GNB: Gaussian Naive Bayes, u Receiver Operating Characteristic - Area Under the Curve.