Dilek Yildiz , International Institute for Applied Systems Analysis
Guy J. Abel, University Of Hong Kong
The migration statistics for the EU member states are more detailed and frequent, and of better quality than most of the countries worldwide. However, while Eurostat provides information on migrant stocks and migration flows between member states, the age, sex and educational attainment distributions of migration flows over time are still not complete. In this paper, we propose a methodology to estimate the migration flow proportions by age, sex and educational attainment. This ongoing work is based on previous research which estimated immigration and emigration flow proportions globally. The previous work employed random forest models to estimate the sex specific proportions by age and education. In this paper, we introduce the methodology for estimating immigration and emigration flow proportions using random forest models, and present how we envisage to extend our methodology to estimate bilateral flow proportions by age, sex and education.
Presented in Session 2. Machine Learning Approaches for Population Research