Table 4.

The predictive performance of machine learning-based models for the PE and its subtypes.

ML Model	Type of PE	TPR (%)	FNR (%)	PPV (%)	FDR (%)	Accuracy (%)	AUC Value	Precision	Recall	F1 Score
DT	All PE	94.1	5.9	91.4	8.6	92.8	0.93	0.91	0.94	0.93
	EO-PE	92.9	7.1	75	25	94.1	0.95	0.93	0.75	0.86
	LO-PE	66.7	33.3	92.9	7.1	88.2	0.80	0.93	0.93	0.93
	Moderate PE	75	25	91.7	8.3	82.4	0.80	0.85	0.92	0.88
	Severe PE	82.1	17.9	44.4	55.6	79.4	0.70	0.67	0.44	0.53
NB	All PE	96.3	3.7	96.4	3.6	98.6	0.98	0.96	0.96	0.98
	EO-PE	96.4	3.6	80	20	91.2	0.88	0.67	0.80	0.73
	LO-PE	33.3	66.7	87.1	12.9	85.3	0.72	0.96	0.87	0.92
	Moderate PE	25	75	79.3	20.7	73.5	0.68	0.88	0.79	0.84
	Severe PE	89.3	10.7	50	50	82.4	0.67	0.50	0.50	0.50
SVM	All PE	70.6	29.4	77.8	22.2	85.5	0.98	0.71	0.78	0.88
	EO-PE	96.4	3.6	80	20	91.2	0.91	0.67	0.80	0.73
	LO-PE	33.3	66.7	86.7	13.3	82.4	0.76	0.93	0.87	0.90
	Moderate PE	37.5	62.5	80.8	19.2	70.6	0.49	0.81	0.81	0.81
	Severe PE	85.7	14.3	20	80	73.5	0.64	0.17	0.20	0.18
RF	All PE	94.1	5.9	91.4	8.6	92.8	0.94	0.91	0.94	0.93
	EO-PE	92.9	7.1	71.4	28.6	91.2	0.94	0.83	0.71	0.77
	LO-PE	66.7	33.3	92.9	7.1	88.2	0.84	0.93	0.93	0.93
	Moderate PE	87.5	12.5	94.4	5.6	70.6	0.79	0.65	0.94	0.77
	Severe PE	85.7	14.3	33.3	66.7	76.5	0.76	0.33	0.33	0.33

Table 4 legend: All PE—all types of preeclampsia; EO-PE—early- onset preeclampsia; LO-PE—late-onset preeclampsia; ML—machine learning; DT—decision trees; NB—naïve Bayes; SVM—support vector machine; RF—random forest; TPR—true positive rate; FNR—false negative rate; PPV—positive predictive value; FDR—false detection rate; AUC—area under the curve.