Ensemble Classifier for Eye State Classification using EEG Signals
This addresses the need for improved classification methods in BCI applications, but it is incremental as it builds on existing techniques.
The paper tackled EEG-based eye state classification for brain-computer interfaces by comparing SVM, HMM, and RBF methods and proposing an ensemble model using random forest and Kstar with a voting algorithm, achieving 97.27% accuracy and 0.13 MAE.
The growing importance and utilization of measuring brain waves (e.g. EEG signals of eye state) in brain-computer interface (BCI) applications highlighted the need for suitable classification methods. In this paper, a comparison between three of well-known classification methods (i.e. support vector machine (SVM), hidden Markov map (HMM), and radial basis function (RBF)) for EEG based eye state classification was achieved. Furthermore, a suggested method that is based on ensemble model was tested. The suggested (ensemble system) method based on a voting algorithm with two kernels: random forest (RF) and Kstar classification methods. The performance was tested using three measurement parameters: accuracy, mean absolute error (MAE), and confusion matrix. Results showed that the proposed method outperforms the other tested methods. For instance, the suggested method's performance was 97.27% accuracy and 0.13 MAE.