Analysis of Inter-Provincial Migration Drivers in Burkina Faso Using Machine Learning Approach

This study aimed to quantify the impacts of migration drivers on out-migration and in-migration in Burkina Faso. The study covered the 45 provinces of Burkina Faso and the three climatic zones. The approach was based on the use of machine learning specifically XGBoost Regression, Random Forest, Neural Network, Linear Regression, Lasso Regression, SGD Regression and Ridge Regression. The results showed that 1) the out-migration is influenced by the population, minimal temperature, masculinity, population density, rural population and the maximal temperature and 2) the in-migration is influenced by the rural population. In perspectives, the approach will be used to predict the migration flow in Burkina Faso under different different scenarios.