Predicting Global Nuclear Energy Suitability Using Machine Learning Analysis

Abstract

This study addresses the global reliance on non-renewable resources for electricity production in a society seeking cleaner, more reliable alternatives. The research goal is to predict the most suitable countries for new nuclear power plants based on climate factors, existing plants, and the electricity they already generate. We did this by testing various machine learning models, including Logistic Regression, Random Forest, Support Vector Machines (SVM), K-Nearest Neighbor (KNN), Decision Tree, and Gradient Boosting, to identify the most efficient classification algorithm. Our results indicate that the KNN model achieved the highest performance metrics. Using our KNN model, Burundi was predicted as the most suitable country for future nuclear development.