Malaria remains a pressing public health challenge in Uganda.
This study investigates the spatial and predictive modeling of
malaria incidence across Uganda's districts.
A spatial weights matrix defined neighborhood relationships,
followed by an Ordinary Least Squares (OLS) regression.
Spatial dependence was addressed using Spatial Lag (SLM)
and Spatial Error (SEM) models, guided by Lagrange
Multiplier tests. In contrast, Geographically Weighted
Regression (GWR) was used to evaluate local risk factor
variations. Advanced models, including Random Forest (RF),
XGBoost, Geographically Weighted Random Forest (GWRF),
and Spatial XGBoost, enhanced accuracy by capturing non-
linear patterns