Social mining for the classification of mental illnesses in public forums

• Autor(a) principal: Ferreira, Rodrigo Miguel Maia
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/38990
• topic: Data mining; Machine learning; Natural language processing; Mental health

The increasing amount of mental health issues is one of the biggest adversities that we face nowadays as a society, and the traditional assistance methods often fail to help those in need. In this work, we implement and evaluate the performance of a screening tool that may complement some of the traditional methods’ weaknesses, by signalling subjects at risk of developing mental illnesses, that could benefit from receiving medical assistance. This tool is based on machine learning, and it detects individuals at risk using their publicly available data from the social network Reddit. This work was based on our participation in tasks 1 and 2 of the 2022 edition of CLEF eRisk, with the goal of detecting subjects at risk of pathological gambling and depression respectively, where we had a special focus on the use and comparison of different text vectorization methods. Despite the fact that the initial results obtained at the event were far from those desired, with some tweaks and additional experiments, we managed to improve them, achieving final F1-scores of 0.886 and 0.653 for the best models of tasks 1 and 2 respectively.