Analyses of relationships between life satisfaction and environmental conditions with machine learning applications

The aim of this thesis is to analyze relationships between environmental factors and life satisfaction. In this regard, individual level data from the fourth European Quality of Life Survey (EQLS) conducted by Eurofound in 2016 is utilized. Mixed-effects multilevel models for ordered responses are implemented. Multilevel analyses imply that less noise level, easy access to green space, and easy access to recycling services have positive relations with life satisfaction of individuals. On the other hand, bad air quality and life satisfaction have a negative relationship. Besides, heavy traffic does not indicate significant relationship with life satisfaction of individuals. The machine learning algorithms such as SVM, KNN, WKNN, XGBoost, GBM, bagging, a decision tree for ordinal responses, and ordinal forest are employed for further analysis. However, machine learning algorithms do not indicate an association between environmental factors and life satisfaction. Standard of living, income, mental wellbeing index, age and health are determined as the most important variables for life satisfaction.