Discrete clusters formulation through the exploitation of optimized k-modes algorithm for hypotheses validation in social work research: the case of greek social workers working with refugees

Abstract

This article focuses on the results of self-funded quantitative research conducted by social workers working in the “refugee” crisis and social services in Greece (1). The research, among other findings, argues that front-line professionals possess specific characteristics regarding their working profile. Statistical methods in the research performed significance tests to validate the initial hypotheses concerning the correlation between dataset variables. On the contrary of this concept, in this work, we present an alternative approach for validating initial hypotheses through the exploitation of clustering algorithms. Toward that goal, we evaluated several frequently used clustering algorithms regarding their efficiency in feature selection processes, and we finally propose a modified k-Modes algorithm for efficient feature subset selection.