BMC Medical Informatics and Decision Making

Table 1 Comparative results of methods with different window sizes

From: Representation learning for clinical time series prediction tasks in electronic health records

	30 Days	60 Days	90 Days	180 Days
RNN-DAE	0.783	0.783	0.783	0.783
RNN-AE	0.755	0.755	0.755	0.755
SDAs	0.488	0.738	0.741	0.755
Hand	0.525	0.584	0.586	0.608
PCA	0.504	0.555	0.555	0.602
GMM	0.536	0.595	0.594	0.607
k-means	0.569	0.568	0.568	0.628

The performance is measured by AUC (i.e., the area under the ROC curve).

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com