Machine Learning Models for Predicting Solar Power Potential and Energy Efficiency in some underserved localities of the Congo Basin Region (Proposals Track)
Jean de Dieu NGUIMFACK NDONGMO (The University of Bamenda); Adelaide Nicole KENGNOU TELEM (University of Buea); Reeves MELI FOKENG (The University of Bamenda)
Abstract
This research proposal aims to develop machine learning models—specifically Linear Regression, Long Short-Term Memory, and Convolutional Neural Networks—to accurately predict solar power potential and energy efficiency in selected underserved localities within the Congo Basin. The Congo Basin, recognized for its ecological significance as the world's second-largest tropical rainforest, faces severe energy access challenges, especially in rural communities that rely on traditional biomass for heating and cooking. This dependency intensifies deforestation and contributes to climate change through increased greenhouse gas emissions. Despite substantial solar energy potential of the region, access to clean and renewable energy sources remains limited. The study will compare the models based on accuracy, reliability, training times, and memory usage, generating actionable insights for development agencies and local stakeholders. By enabling informed, data-driven decisions regarding sustainable energy solutions, this work intends to facilitate a transition from traditional biomass to renewable energy sources, ultimately contributing to both environmental conservation and improved quality of life for local populations.