Spatial Epidemiology and Remote Sensing-Informed Prediction Modelling of Malaria Transmission Risk in North West Nigeria Using Machine Learning Approaches

Spatially heterogeneous malaria transmission in North West Nigeria reflects complex interactions among climatic, ecological, demographic, and entomological factors that conventional epidemiological approaches cannot adequately model for precision public health targeting. This research applies spatial epidemiology and machine learning to develop a high-resolution malaria transmission risk prediction model for Kano, Katsina, and Zamfara States in North West Nigeria. A multi-source data integration design will combine malaria case data from health facility records of 300 sites over five years, Sentinel-2 and MODIS satellite imagery for land cover and normalised difference vegetation index, climate reanalysis data, Anopheles mosquito density estimates, household survey data, and demographic variables. Random forest, gradient boosting, and deep learning convolutional neural network models will be trained and compared for 1-kilometre-resolution malaria risk prediction. Spatial autocorrelation analysis using Moran's I and LISA cluster detection will characterise transmission hotspot dynamics. Temporal prediction incorporating seasonal rainfall patterns will enable three-month transmission forecasting. Cross-validation at 20% held-out sites will assess generalisability. Model outputs will be visualised in a GIS-linked decision support dashboard accessible to state malaria control programmes. Original contributions include the first machine learning-based malaria prediction platform for North West Nigeria, integration of remote sensing at sub-kilometre resolution with entomological data, and a theoretical extension of spatial risk modelling frameworks to West African transmission heterogeneity. Findings will directly optimise indoor residual spraying and bednet distribution targeting in North West Nigeria. Keywords: spatial epidemiology, malaria prediction, machine learning, remote sensing, North West Nigeria