Prediction of emergency department revisits among child and youth mental health outpatients using deep learning techniques

Table 2 Performance metrics of GNN, RNN, conventional machine learningand logistic regression classification models

Model type	F1 scores (SD)	Precision	Recall	AUROC	Accuracy (%)	PPV (%)	NPV (%)
GNN	.6502 (.0356)	.6369	.6654	.7022	64.43	64.24	65.37
RNN	.5991 (.0430)	.6502	.5566	.6611	62.82	64.76	61.03
XGBoost	.6032 (.0437)	.6414	.5707	.6919	62.89	65.49	62.08
Decision tree	.5779 (.0462)	.5838	.5736	.5908	58.64	59.36	58.76
Logistic regression	.5828 (.0415)	.6166	.5538	.6455	60.58	62.22	59.84

AUROC Area under the Receiver Operating Characteristic curve, SD standard deviation, PPV positive predictive value, NPV negative predictive value

ISSN: 1472-6947