Early detection of eating disorders through machine learning techniques

Abstract

In this work we analyze the relationships between nutrition, health status and well-being of the individual in evolutionary age, not only in consideration of the high prevalence of excess weight and the early appearance of metabolic pathologies, but also due to the significant presence of Eating Disorders (EDs). EDs, in fact, continue to be underdiagnosed by pediatric professionals and many adolescents go untreated, do not recover or reach only partial recovery. We have observed the situation of young people at an Italian High School regarding EDs by carrying out a statistical survey on the students in relation to dietary habits, attitudes towards food and physical activity. Finally, the collected data have been analyzed through statistical and machine learning techniques.