Inteligência artificial aplicada no mapeamento de sintomas e tratamento de depressão, ansiedade e estresse

• Autor(a) principal: Delgado, Sabrinna
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da Uninove
• Texto Completo: http://bibliotecatede.uninove.br/handle/tede/3243
• topic: transtornos mentais; DASS-21; mapeamento de sintomas; inteligência artificial; mineração de dados; lógica fuzzy; mental disorders; DASS-21; symptoms mapping; artificial intelligence; data mining; fuzzy logic; CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO

Every year the number of people in the world affected by mental disorders (MD) increases, among which are depression, anxiety and stress that have been more common and that are usually related to the modern lifestyle. The first two belong to the group of the main diseases of the 21st century and can lead to serious consequences, such as suicide. According to the Pan World Health Organization (WHO), depression impacts the daily lives of more than 300 million people, being considered one of the most important diseases in the world. Additionally, an estimated 12 billion workdays are lost annually worldwide due to depression and anxiety, impacting nearly a trillion dollars on the global economy. TM treatment may include, in addition to medication and psychotherapies, which are essential, the use of technological resources, such as Artificial Intelligence (AI) to indicate therapies and personalized care. In the literature, there are several AI approaches applied in the context of MT, but it is very common that they are focused on aiding the diagnosis. This research proposes an AI method for mapping symptoms and helping to treat depression, anxiety and stress. First, data mining (DM) techniques are applied to generate rules that, in addition to mapping the symptoms, represent knowledge about a database containing data from 242 patients, collected from a test called DASS-21 (Depression, Anxiety and StressScale). Then, the generated set of rules is used to compose a Fuzzy Inference System (FIS) capable of making predictions about MDs based on the main symptoms and some personal data of the patient. The high hit rates in the DM tasks (above 90%) indicating the existence of consistent patterns and the results produced by the FIS demonstrate that the proposed method can help health professionals in the rapid prediction of symptoms of depression, anxiety and stress, in outpatient screening and in emergency care. It can also be useful for a better association of symptoms, therapeutic proposals and even investigations of other diseases not related to mental health, providing differential diagnoses and treatments.