Table 1.
Classifying CVD based on self-reported health conditions.
Fig 1.
Co-occurrence matrix of cardiovascular diseases, illustrating the frequency of co-existing conditions in the dataset.
Fig 2.
Data preprocessing workflow from raw data to modeling-ready dataset.
Fig 3.
RFE plot for feature selection.
Table 2.
Descriptive summary of final dataset with p-values.
Table 3.
Performance ML models.
Fig 4.
ROC curve comparison for all ML models.
Fig 5.
LIME explanations for selected individual predictions, showing features supporting or contradicting the predictions.
Fig 6.
SHAP beeswarm plot illustrating the distribution and impact of feature values on model predictions.
Fig 7.
SHAP dependence for Total_Cholesterol on the test set.