iMind: Uma ferramenta inteligente para suporte de compreensão de conteúdo

• Autor(a) principal: Gomes, Miguel Duarte Sousa
• 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/10400.22/22237
• topic: Biometrics measurement; Cognitive load; Content comprehension; Eyetracking; Heart rate variability; Machine learning

Usually while reading, content comprehension difficulty affects individual performance. Comprehension difficulties, e. g., could lead to a slow learning process, lower work quality, and inefficient decision-making. This thesis introduces an intelligent tool called “iMind” which uses wearable devices (e.g., smartwatches) to evaluate user comprehension difficulties and engagement levels while reading digital content. Comprehension difficulty can occur when there are not enough mental resources available for mental processing. The mental resource for mental processing is the cognitive load (CL). Fluctuations of CL lead to physiological manifestation of the autonomic nervous system (ANS), which can be measured by wearables, like smartwatches. ANS manifestations are, e. g., an increase in heart rate. With low-cost eye trackers, it is possible to correlate content regions to the measurements of ANS manifestation. In this sense, iMind uses a smartwatch and an eye tracker to identify comprehension difficulty at content regions level (where the user is looking). The tool uses machine learning techniques to classify content regions as difficult or non-difficult based on biometric and non-biometric features. The tool classified regions with a 75% accuracy and 80% f-score with Linear regression (LR). With the classified regions, it will be possible, in the future, to create contextual support for the reader in real-time by, e.g., translating the sentences that induced comprehension difficulty.