Table 2 Comparison of model performances between the proposed model and the conventional machine learning models.

From: Predicting the need for intubation within 3 h in the neonatal intensive care unit using a multimodal deep neural network

Model

Mean (SD)

AUROC

F1-score

Sensitivity, %

Specificity, %

Accuracy, %

LR

0.886 (0.081)

0.839 (0.093)

83.4 (6.1)

83.2 (13.3)

83.3 (9.8)

XGBoost

0.853 (0.069)

0.810 (0.094)

79.5 (7.7)

80.5 (11.8)

80.5 (9.9)

SVM

0.890 (0.052)

0.882 (0.052)

82.7 (6.0)

89.7 (7.4)

88.0 (5.4)

MDNN

0.917 (0.042)

0.884 (0.048)

85.2 (9.3)

89.2 (6.8)

88.2 (5.0)

  1. LR linear regression, XGBoost extreme gradient boosting decision tree, SVM support vector machine, MDNN multimodal deep neural network, AUROC area under the curve of receiver operating characteristics.
Back to article page