Dissertação
Extraction of structural and semantic features for the identification of Psychosis in European Portuguese EVALUATED
Psychosis is a mental condition that affects the subject’s behavior and perception of the world, impairing both his cognitive and speech capabilities. These impairments, when associated with the stigma that mental disorders carry in society, promote the subject’s disconnection from reality and society. Psychosis, and other mental disorders, lack efficient and accurate diagnostic tools, relying mostly on self-reports from patients, their families, and specialized clinicians. Studies identified by us, typically focus on the identification or prediction of psychosis through surface-level analysis of the subject’s speech targeting audio, time, and paucity features. Recent studies have started focusing on high-level and complex language analysis such as semantics, structure, and content. A very limited number of studies have targeted the Portuguese language. To the best of our knowledge, no study has focused on structural or semantic features in European Portuguese, which is our main objective. Our work also aimed at expanding the First European Corpus for Psychosis Identification with more subjects and a more diverse control group, to better understand the impact that such alterations have on the results achieved by the previous studies. Results obtained support future developments of models for the identification of psychosis, that rely on structure, coherence, and content analysis of discourse. However, it also suggested that further developments in Natural Language Processing techniques, especially for European Portuguese, are required for the progression and improvement of the results obtained. The models developed by our solution were verified to be reliable and robust.
novembro 24, 2022, 18:0
Publicação
Obra sujeita a Direitos de Autor
Orientação
ORIENTADOR
Helena Sofia Andrade Nunes Pereira Pinto
Departamento de Engenharia Informática (DEI)
Professor Auxiliar
