Predictive model and risk analysis for coronary heart disease in people living with HIV using machine learning

Table 2 Performance of prediction models generated by the seven ML algorithms

Models	AUC	AUC 95% CI		Recall	SP	ACC	F1	PPV	NPV
Models	AUC	Lower bound	Upper bound	Recall	SP	ACC	F1	PPV	NPV
LightGBM	0.849	0.814	0.883	0.721	0.858	0.757	0.238	0.143	0.989
XGBoost	0.819	0.779	0.859	0.698	0.796	0.784	0.176	0.101	0.988
AdaBoost	0.787	0.742	0.832	0.558	0.889	0.961	0.225	0.141	0.984
Multilayer Perceptron	0.840	0.804	0.876	0.744	0.822	0.967	0.207	0.120	0.990
Decision Tree	0.753	0.704	0.803	0.953	0.389	0.407	0.093	0.049	0.996
Support Vector Machine	0.828	0.790	0.866	0.744	0.793	0.675	0.185	0.106	0.990
Lasso-Logistic	0.843	0.807	0.878	0.884	0.682	0.744	0.153	0.084	0.994

ISSN: 1472-6947