Fig. 3

A an importance matrix plot of the LightGBM model depicting the significance of each variable in anticipating CHD risk in PLHIV. B SHAP summary plot of the top eight clinical attributes of the LightGBM model. Each point represents the SHAP value of a specific feature on a data point, indicating the magnitude and direction of that feature’s impact on the model’s predictive outcome. Red points denote high feature values with a positive incremental effect on the prediction; blue points denote low feature values with a negative decremental effect. Features are ranked from top to bottom by their average impact, highlighting their relative importance in the model’s decision-making process. Glu, glucose; Scr, serum creatinine; I-Bil, indirect bilirubin; AMY, amylase; SUA, serum uric acid