Robust SNP-based prediction of rheumatoid arthritis through machine-learning-optimized polygenic risk score

Table 1 Predictive performance of the 9 selected SNPs in a fivefold cross-validation of the Training dataset

Dataset	Evaluation metric	Machine learning models
Dataset	Evaluation metric	Logistic regression	Naïve bayes	Random forest	XGBoost	SVM RBF
Training set Cross-validation	Mean AUC	0.992	0.990	0.994	0.994	0.992
	Mean Sensitivity	0.968	0.975	0.975	0.973	0.968
	Mean Specificity	0.963	0.956	0.962	0.963	0.965
	Mean accuracy	0.966	0.966	0.968	0.968	0.966
	Mean average precision (PR-AUC)	0.979	0.973	0.980	0.981	0.968

ISSN: 1479-5876