Table 4.

ML applications in SA research.

Reference	ML application	Input	ML technique	Main output
(74)	Complication prediction	Clinical parameters	LR	AUC = 0.924, Se = 0.922, Sp = 0.733, PPV = 84.7% and NPV = 85.4%
(77)	Complication prediction	Genomics data	SVM	AUC = 0.71, Ac = 67%, Se = 86% and Sp = 43%
(78)	Complication prediction	Proteomics data	DT	AUC = 1, Ac = 100%, Se = 100%, Sp = 100%, Kappa = 1, PPV = 1 and NPV = 1 for recurrent SA with prethrombotic state
(79)	Complication prediction	Clinical parameters	RF	AUC = 0.99, Ac = 0.99, Pr = 0.99, Re = 0.99, F1 = 0.99
(80)	Complication prediction	Electronic health records	SC	AUC = 0.909 and Ac = 89.7% for live birth

ML, machine learning; SA, spontaneous abortion; LR, logistic regression; SVM, support vector machines; DT, decision tree; RF, random forest; SC, sparse coding; AUC, area under the receiver operating characteristic curve; Se, sensitivity; Sp, specificity; PPV, positive predictive value; NPV, negative predictive value; Ac, accuracy; Pr, precision; Re, recall.