Skip to main content
. 2023 Dec 22;25:e48244. doi: 10.2196/48244

Table 2.

Results predicted by the proposed model for different time windows.

Window and feature Precision, mean (SD) Sensitivity, mean (SD) Specificity, mean (SD) F1-score, mean (SD) AUROCa, mean (SD) AUPRCb, mean (SD)
12-hour window






Statistical features 0.53 (0.00) 0.65 (0.02) 0.65 (0.02) 0.35 (0.01) 0.65 (0.02) 0.26 (0.05)

Similarity features 0.54 (0.00) 0.71 (0.01) 0.71 (0.01) 0.41 (0.01) 0.71 (0.01) 0.38 (0.06)

All statistical features 0.54 (0.00) 0.73 (0.02) 0.73 (0.02) 0.45 (0.01) 0.73 (0.02) 0.40 (0.05)
24-hour window






Statistical features 0.53 (0.00) 0.69 (0.03) 0.69 (0.03) 0.41 (0.01) 0.69 (0.03) 0.34 (0.09)

Similarity features 0.57 (0.00) 0.84 (0.02) 0.84 (0.02) 0.56 (0.00) 0.84 (0.02) 0.52 (0.06)

All statistical features 0.68 (0.04) 0.90 (0.03) 0.90 (0.04) 0.72 (0.04) 0.86 (0.02) 0.58 (0.07)

aAUROC: area under the receiver operating characteristic curve.

bAUPRC: area under the precision-recall curve.