Conflict susceptibility mapping methodology and data experiments

This research presents an innovative approach and methodology to conflict susceptibility mapping by integrating machine learning technologies with geospatial data analysis. Utilizing public datasets from Somalia, the study experiments with applying machine learning models such as random forest classifier, support vector machine classifier, and gradient boosting classifier models to predict areas susceptible to conflict. The methodology includes data preprocessing, model training, execution, and validation, employing various software and machine learning techniques. The random forest classifier emerged as the most accurate model through experiments with the machine learning models, demonstrating the potential of using machine learning to enhance our understanding of conflict dynamics. The study highlights the critical role of selecting appropriate conditioning factors and the need to continuously refine methodologies to improve prediction accuracy. By providing a practical method for conflict susceptibility mapping, this research contributes to the broader field of peace and security research, which directly contributes to Sustainable Development Goal 16 and explores the potential of using machine learning to support peace, justice, and strong institutions, and contribute to global peace and security.