An Adaptive and Efficient Method for Detecting First Signs of Depression with Information from the Social Web

Abstract

Depression is one of the most prevalent mental disorders in the world. At present, there are more than 264 million people of all ages suffering depression and close to 800000 ends up in suicide (World Health Organization (Jannuary 2020)). The early recognition of depression can lead to timely treatment and save lives. In this context, the use of information from social media platforms can be a valuable resource for the early detection of depression. Previously, we presented k-TVT, a method able to set the level of urgency for the detection of this mental disorder. This adaptive method considers the variation of the vocabulary along the time step line for representing the documents. The results obtained with k-TVT using publicly available data sets demonstrated its flexibility and effectiveness over state-of-the-art methods. In this extended work, we confirm the previous conclusions with a more elaborated analysis of results.