Classification of Left/Right Hand Movement from EEG Signal by Empirical Mode Decomposition with Ensemble of classifiers

Electroencephalography (EEG) is a brain wave signal that plays a significant role in the neural control interface. EEG signals have many nonlinear dynamic properties; because of these properties, handling EEG signals for modeling with standard linear approaches will lead to inconsistent results. EEG has a significant importance in the brain-computer interface (BCI). BCI has demonstrated a great deal of potential to benefit the human way of life. Additionally, its application targets people with disabilities. BCI extracts the informative featuresfrom acquired brain signals and generates signals knowing these features, which can control external devices. EEG signals can show improvement in classifying motor imagery tasks. In this paper, we introduce a new method for extracting time-frequency-related properties on different scales using empirical mode decomposition (EMD) and classifying the left and right hand movements of the subject. EMD is a signal processing technique that decomposes a signal into different frequency bands of the same length. We use an ensemble of classifiers with bootstrapped data for classification. The proposed method perform better than the previous method [1] in terms of accuracy and F1-score. The empirical results show that the proposed method has an average accuracy of 97.30% and an average F1 score of 96.35%.