Muthono19/Predicting-aflatoxins-levels-in-maize-using-machine-learning

The R-Scripts and data uploaded in this repository is part of IITA project on mapping aflatoxins level in maize as part of Ukama Ustawi (UU): Diversification for Resilient Agri-food in East and Southern Africa CGIAR initiative # Predicting-aflatoxins-levels-in-maize-using-machine-learning The R-Scripts and data uploaded in this repository is part of IITA project on mapping aflatoxins level in maize as part of Ukama Ustawi (UU): Diversification for Resilient Agri-food in East and Southern Africa(cgiar.org) CGIAR initiative. Climate change and variability (CCV) has major impact on agriculture in East and Southern Africa (ESA) region. Smallholder subsistence farmers practicing rain-fed agriculture in Sub-Sahara Africa (SSA) are the most vulnerable to the impacts of climate change and variability. CCV lead to novel outbreaks and/or expansion of geographical range of pests and diseases. Mycotoxins contamination of major staple cereals like maize is a major health hazard to humans and livestock. Two of the most important mycotoxigenic fungi associated with maize, fruits and nuts are Aspergillus flavus, which produces aflatoxins, and Fusarium verticillioides, which produces fumonisins. Contamination of grains by aflatoxins is influenced by climate (temperature and humidity), soil type and nutrients and crop management practices (tillage method, time of planting and crop rotation regime) . Understanding the spatial variation of the aflatoxins prevalence and its driving factor is important for spatial targeting control measures such as Aflasafe®. Previous attempt to model the drivers of aflatoxins has largely ignored the crop management practices , yet they are significant drivers of contamination rates. Nearby farms with similar weather and soils can have contrasting concentration of Aflatoxins due to management practices such as timing of planting. Traditional species distribution models use continuous environmental variables (climate, soil and topography), however recent methodological advances have developed methods of incorporating categorical management practices in spatial models