Liu X, Jiang J, Wei L, et al. - Using machine learning (ML) methods, three predictive models were constructed for all-cause mortality in coronary artery disease (CAD) cases with atrial fibrillation (AF), and after assessing performance of all models, the regularization logistic regression model was finally recommended to be employed in clinical practice.

• This study involved 2,037 CAD patients with AF, and three ML methods were employed, including the regularization logistic regression, random forest, and support vector machines.

• Models were developed incorporating 24 variables, and an area under the curve (AUC) of 0.732, 0.728, and 0.712 were generated by the regularization logistic regression, random forest, and support vector machines model, respectively.

• Among the three models, the highest AUC value (0.732), specificity (0.699 vs 0.663 vs 0.668), and accuracy (0.936 vs 0.935 vs 0.935) were offered by the regularization logistic regression model.

• However, the receiver operating characteristic curve of the three models did not demonstrate statistical differences.