EEG eye state based classification using supervised learning

With the widely used of brain-computer interfacing (BCI) technology, the need for fast and accurate classification. The classifier system developed in the current study improved upon the previous study developed by Kadir Sabanci & Murat Koklu (2015) [24] for detecting whether an individual's eyes are open or closed on EEG recording. In this paper, removed the outlier (i.e. error transmission) from the database which is obtained from UCI machine learning repository and classify the data by using K-Nearest Neighbour and Multilayer Perceptron neural network. The obtained success classification methods were compared. All classification method performed in the weka tool by default parameters. Finally, the highest classification success rate obtained with the kNN 91.82% on 1 nearest neighbourhood. However, the highest classification success rate obtained with the Multilayer perceptron 72.42% on the number of neurons in the hidden layer was equal to 20.