Fig 1.
Subject selection flowchart.
Fig 2.
Architecture of model construction.
BP, BP neural network; DL, deep learning; LR, logistic regression; RF, random forest; XGB, XGBoost.
Table 1.
Performance Evaluation of Three Models Before and After SMOTE Resampling.
Fig 3.
Comparative analysis of evaluation metrics for three models before and after SMOTE resampling.
Table 2.
The specific variables and their correlation levels.
Table 3.
Selected baseline characteristics of participants.
Table 4.
Performance of each model.
Fig 4.
Performance curves of various models.
Fig 5.
Bar chart and honeycomb plot of the top 15 variable features in BP neural network model.
Fig 6.
Bar chart and honeycomb plot of the top 15 variables in Logistic Regression model.
Fig 7.
Bar chart and honeycomb plot of the top 15 variables in the XGBoost model.
Fig 8.
Bar chart and honeycomb plot of the top 15 variables in the Random Forest model.
Fig 9.
Explanation of variables in varimp function of Deep Learning.