Crop Protection and Productivity through the Adoption of Machine Learning Algorithms in Sub-Saharan Africa: A Systematic Review

Abstract

The economy of Sub-Saharan Africa depends heavily on agriculture, facing challenges like inefficient resource use, food insecurity, and climate shocks. This paper evaluates how machine learning (ML) could transform agricultural practices. We reviewed 88 studies (2010-2025) on ML applications like yield prediction, precision agriculture, and disease detection in countries like Rwanda, Nigeria, and Kenya. Systems like SMART-Crop Yield Prediction System predict maize yields with ≤0.177% inaccuracy using soil and meteorological data. CropGuard diagnoses crop diseases with 97% reli-ability using image recognition. ML can improve harvests by optimizing irrigation, cutting fertilizer waste by 30%, and enabling early pest action. Challenges include low technical literacy, erratic electricity, limited internet, and cultural resistance to adopting new tech. Scaling ML could strengthen food security, cut post-harvest losses, and em-power smallholder farmers. Success requires collaboration between governments, tech developers, and local communities. Priorities include developing farmer-friendly ML tools, funding rural infrastructure, and training initiatives. Sub-Saharan Africa may use ML for climate-resilient farms, stabilizing food supplies, and improving livelihoods by aligning tech with local needs, paving the way for sustainable agricultural growth.