Depression Recognition in Social Media based on Symptoms' Detection

Resumen

Depression is a common mental disorder that affects millions of people around the world. Recently, several methods have been proposed that detect people suffering from depression by analyzing their language patterns in social media. These methods show competitive results, but most of them are opaque and lack of explainability. Motivated by these problems, and inspired by the questionnaires used by health professionals for its diagnosis, in this paper we propose an approach for the detection of depression based on the identification and accumulation of evidence of symptoms through the users’ posts. Results in a benchmark collection are encouraging, as they show a competitive performance with respect to state-ofthe- art methods. Furthermore, taking advantage of the approach’s properties, we outline what could be a support tool for healthcare professionals for analyzing and monitoring depression behaviors in social networks.