Unifying Application and Adaptative Patterns for Confused student EEG Signals Classification

During COVID-19 pandemic, online education has become a crucial educational tool. While it offers numerous benefits, it does not have face-to-face interactions, making it challenging to assess students' comprehension levels and detect confusion. This research utilizes electroencephalogram (EEG) data to identify confusion in students using the MOOC platform. Current ways to detect confusion primarily concentrate on optimizing models, and the study of feature engineering is relatively limited. Brain fog, or confusion, is a significant cause of low performance in learning and other cognitive tasks. The ability to detect confusion in world time is a complex and essential matter that has numerous applications, including online education and driver fatigue detection. The paper employsBidirectional LSTM Recurrent Neural Networks to acknowledge a student’s difficulties and confusion while watching course videos on the internet, using EEG data. The study reveals to us that the Bidirectional LSTM model outperforms other machine learning methods anddemonstrates remarkable stability in cross-validation. The study achieves an accuracy rate of % in predicting student confusion. Additionally, the research highlights that the wave (gamma) of the EEG signal is the most critical feature in detecting brain confusion. These findings underscore the potential of machine learning to model and comprehend brain function.