Optimal Machine Learning Approach for EEG Eye-State Classification in Metaverse Environment

In this paper, we delve into the application of logistic regression to classify eye states from EEG data tailored for inte-gration within Metaverse interfaces. Using Sequential Backward Selection, we meticulously optimize feature subsets across varying regularization strengths to enhance model accuracy. Our study delineates the delicate balance between model complexity and performance, uncovering an optimal feature count that maximizes accuracy. We highlight the trade-offs between precision and recall, underscoring their implications for the development of sophisticated, real-time classifiers. Our contributions are poised to advance neuro-technological applications in virtual environments, offering a pathway to more intuitive and responsive Metaverse interactions.