SHAP-enhanced machine learning identifies modifiable obesity predictors across adolescent weight groups: A 2021 YRBSS analysis
Fig 2
Beeswarm plot of SHAP values for the top 15 predictors of adolescent obesity in the XGBoost model. (Note: Each point represents an individual participant. The x-axis shows the SHAP value (impact on log-odds of obesity), where positive values indicate increased risk and negative values indicate reduced risk. The y-axis lists predictors ranked by overall importance, and point colour reflects the feature value (yellow = high, purple = low). The least influential variable (Cigarettes) was excluded).