A machine learning-based assessment of subjective quality of life

This paper evaluates subjective quality of life using various Machine Learning (ML) techniques. Utilizing a survey conducted in Quito's historical downtown and adopting subjective quality of life as a theoretical framework, we applied ML regression techniques (ordinal logistic regression, random forests and support vector machines) to identify satisfactions with different domains of life influencing overall satisfaction with life. Our findings indicate that satisfaction with work, satisfaction with household conditions and satisfaction with health explain overall life satisfaction. Additionally, it was observed that high values of these satisfactions are evenly distributed across space. The applied techniques can be easily replicated for diverse urban contexts, expanding the use of ML methods within the Social Sciences. Assessing satisfaction with life of urban residents is critical to support better urban planning. Policy-making should promote satisfaction of different domains of life by recognizing the complexity and multidimensionality of the factors that influence the quality of life of human beings.