Print Email Facebook Twitter Language Inquiry for Personalized Mental Health Chatbots Title Language Inquiry for Personalized Mental Health Chatbots Author Mazza, Maria Chiara (TU Delft Technology, Policy and Management) Contributor Rook, L. (mentor) Brazier, F.M. (graduation committee) Degree granting institution Delft University of Technology Programme Management of Technology (MoT) Date 2020-07-08 Abstract Anxiety is one of the most widespread and dangerous mental health disorders in developed and underdeveoped countries, affecting a wide number of college students all over the world. If not correctly treated in time, it might procure irreparable damages in people’s life, leading to drastic consequenses such as depression and suicidal intentions. However, although some students seek for medical consultation, only one quarter of them is able to have access to clinical treatments. At the same time, chatting apps gradually became a new communication trend during the last few years, resulting in the development of a new cutting-edge technology named conversational agents. After several studies, this technology has been found to be a possible solution for the healthcare imparity between demand and supply. With this invention, students might have the possibility to chat with a sort of “online therapist” anywhere and anytime they feel the need, without stigma or judgement barriers. In order to successfully implement these conversational agents, the therapeutic alliance between the doctor and the patient should be recreated as accurately as possible. Personality seems to be an important factor for the success and eventual satisfaction in the whole treatment. The present research – through LIWC software – explores the extent to which students use different linguistic patterns in an expressive writing task depending on their personality and mental health status. This study hypothesized that students sufffering from Generalized Anxiety Disorder (GAD) use different words than mentally stable students, and that their linguistic patterns are further influenced by their behavioral activation or inhibition systems. The main findings were in line with these two hypotheses. Based on the results, both students affected and not by GAD use different words specifically depending on their BAS levels. In conclusion, as predicted by previous researchers, personality is well-reflected through language styles: each student with a specific behavior, mental health characteristic, and even nationality expresses him/her self with different linguistic patterns. Subject PsychologyChatbotConversational AgentPersonalityLanguage InquiryLIWCAnxiety To reference this document use: http://resolver.tudelft.nl/uuid:1d8831ed-7985-4980-91fa-b7c243e126e1 Part of collection Student theses Document type master thesis Rights © 2020 Maria Chiara Mazza Files PDF Maria_Chiara_Mazza_Final_Thesis.pdf 1.91 MB Close viewer /islandora/object/uuid:1d8831ed-7985-4980-91fa-b7c243e126e1/datastream/OBJ/view