Fig. 5

The partial dependence plot of each variable. The partial dependence plots isolate the effects of eight feature variables on the predicted number of VL cases. The X-axis represents the range of values for each feature, while the Y-axis shows the model’s average predicted value for VL cases. The blue curve represents the average predicted values after 50 iterations, and the shaded area indicates the 95% confidence interval for the predictions over these iterations. The red and green line segments represent the range of feature variable values within different clusters, with red indicating high-risk clusters and green indicating low-risk clusters. These ranges are calculated based on the annual average values of variables across all counties within each cluster