Predicting happiness levels of European immigrants and natives: An application of Artificial Neural Network and Ordinal Logistic Regression

Abstract

The main purpose of this paper is to investigate the happiness factors and assess the performance of machine learning techniques on predicting the happiness levels of European immigrants and natives. Two types of machine learning methods, Ordinal Logistic Regression (OLR) and Artificial Neural Network (ANN), are employed for analytical modeling. Our results with a total sample size of 196,724 respondents from nine rounds of the European Social Survey (ESS) indicate that the determinants of happiness for immigrants and natives are significantly inconsistent. Therefore, variables should be specifically selected to predict the happiness levels of these two different groups. The sensitivity analysis shows that satisfaction with life, subjective general health, and the highest level of education are the three most prominent determinants that contribute to the happiness of immigrants and natives. The overall accuracies of OLR and ANN baseline models are >80%. This can be further improved by building models for each individual country. The application of OLR and ANN implies that machine learning algorithms can be a useful tool for predicting happiness levels. The greater knowledge of migration and happiness will allow us to better understand the decision-making processes and construct more effective policies.