GANN: EEG-based Emotion Classification Using Context-Aware Gated Attention Neural Network

Abstract

Electroencephalography (EEG) is a valuable, non-invasive method for monitoring the brain's electrical activity and assessing high-level cognitive processes, including emotions. Despite its benefits, interpreting EEG data is challenging, complex, and prone to human error. This study introduces an advanced machine learning model aimed at improving EEG-based emotion classification by fully utilizing the spatial and temporal characteristics of EEG channels. Our proposed model combines Gated Recurrent Units (GRU) and Attention Mechanisms, significantly enhancing feature learning and model generalization across diverse datasets. Achieving a classification accuracy of 94% on the EEG Brainwave Dataset, our model outperforms existing methods. This work not only refines emotion recognition techniques but also establishes a benchmark for future research in effectively utilizing EEG data.