Utilizing Embedded and Machine Learning Techniques to Classify EEG Eye States

Abstract

Numerous studies focus on epilepsy diseases in order to achieve the detection of eye states and classification systems because of the significance of automatically identifying brain illnesses. Eye condition recognition is essential for biomedical informatics applications like driving detection and smart home device control. Electroencephalogram signals are this problem. In this context, conventional methods and manually derived features are applied in several instances. The extraction of useful features and the choice of appropriate classifiers are difficult problems. This work suggests an ensemble system called "EEG Eye" that employs a new preprocessing stage. In this context, the base classifiers and the most significant classical works are compared to the ensemble approaches in the classification step. A publicly accessible EEG eye state dataset from UCI is used for assessment. At 100%, 100%, 100%, 100%, the maximum accuracy, precision, recall, and F1 are attained.