Prediction of Approved Asylum Seekers in European Countries

This paper is to develop a statistical model that predicts the number of approved asylum seekers using profiles of 32 European countries in a panel data setting. The ordinary least squares (OLS) linear regression, the fixed effects and the random effects models are explored and compared. In addition, the clustering results in 2014 are compared with manually generated clusters. The evaluation results show that the fixed effects model (with “country” and “year” effects) wins out. The k-means clustering and the hierarchical clustering with complete-link have a better performance within a classification on the number of approved asylum seekers. Our study finds the related country profiles, which build a bridge to the study of refugee problem.