Development and performance assessment of novel machine learning models to predict pneumonia after liver transplantation

Table 4 Performance of the six ML models in the testing set

	AUC (95% CI)	Accuracy (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)
LR	0.68 (0.607, 0.743)	0.657 (0.596, 0.713)	0.494 (0.405, 0.595)	0.778 (0.711, 0.84)
SVM	0.676 (0.606, 0.741)	0.646 (0.59, 0.702)	0.62 (0.529, 0.707)	0.67 (0.589, 0.746)
RF	0.781 (0.719, 0.833)	0.736 (0.674, 0.787)	0.632 (0.538, 0.72)	0.813 (0.747, 0.876)
MLP	0.678 (0.611, 0.744)	0.635 (0.579, 0.691)	0.514 (0.423, 0.603)	0.728 (0.651, 0.798)
GBM	0.772 (0.714, 0.827)	0.713 (0.657, 0.77)	0.605 (0.507, 0.697)	0.794 (0.723, 0.856)
XGBoost	0.794 (0.735, 0.84)	0.73 (0.674, 0.781)	0.618 (0.527, 0.705)	0.815 (0.75, 0.872)

Values are expressed as median with interquartile range
LR logistic regression, SVM support vector machine, RF random forest, MLP multilayer perceptron, GBM gradient boosting machine, XGB extreme gradient boosting

ISSN: 1465-993X