Cluster analysis of patients suffering from addictions

Abstract

We investigate the contribution of unsupervised learning to identify patient’s profiles suffering from addictions. We propose a new clustering approach based on coupling b-coloring of graph and Bregman hard clustering algorithm in order to automatically find the number of categories or groups of patients and the “best” representative patients’ profile of each group. The study was carried out in close collaboration with the French co-operative health organization called the “Centre Mutualiste d’Addictologie”, an aftercare centre for addictions. The quantitative data arises from a cohort of seven different aftercare centres for addiction located in France. The study concerns 301 patients suffering from dependence (addictions with psychoactive substances and/or behaviour addictions).