This study explores the connection between perceptions of the quality of social services and people’s

well-being, with a particular emphasis on how people’s well-being affects their perceptions regarding the

quality of services. At the core of this research, we employ a challenging quantitative analysis of a vast EU

database to derive hidden connections and conclusions. Our research’s vision is to use advanced scientific

methods to facilitate decision-making and policy strategies regarding people’s well-being.

The findings of many relevant studies in the literature show the complexity of the relationship between

different variables and mental well-being and especially that, many times, the relationship is not linear.

Therefore, we aim to deeply examine the use of some of today’s most advanced algorithms to identify

patterns, unusual phenomena, and non-linear relationships between the various parameters underlying

the research questions. In particular, many characteristics of the information sometimes manage to fall

outside the capabilities of classical statistical techniques due to different factors, for example, noise (accuracy of measurements), the human nature of answers to survey questions (modeling problem), difficulty

in processing discrete data, issues with scaling data about their “true” weight (normalization), hidden

dependence between parameters, high dimensionality, and more. In this project, we focus on the following

three approaches. First are the more standard clustering and classification algorithms, including unsupervised k-means, spectral clustering, and hierarchical clustering, and supervised learning methods such

as support vector machine and random forest. The second approach is termed questionnaire methods,

where we wish to recognize and exploit the geometry of the underlying data based on learning the most

appropriate metrics or data affinities and applying tree methods and wavelets techniques to analyze the

dataset. The final approach uses the AI-based method, particularly graph neural networks, to define a

more dynamic data representation and analysis model.

The numerical work will be carried out locally at Tel Aviv University. At the same time, the complete

analysis will combine the efforts of all partners, both at Ben-Gurion University and Sapir College. Our

project will form a unique fusion between social and exact sciences research and will introduce a new

interdisciplinary corporation between researchers within different institutions in Israel.