An EEG signals classification system using optimized adaptive neuro-fuzzy inference model based on harmony search algorithm

This paper descries a novel method for classification of human brain activity, such as electroencephalogram (EEG) signals related with motor imagery task using adaptive neuro-fuzzy inference (ANFI) model-based approach. The proposed method was focus on the demonstration of the availability of optimization of ANFI model using Harmony Search algorithm for classifying the motor imagery EEG signals. Before the optimization, the features of the ANFI model classifier are extracted by Hjorth parameters. HS algorithm is sufficiently adaptable to allow incorporation of other ANFI model training techniques like backpropagation, gradient descent method. In order to simulate the proposed method, three types of motor imagery tasks are performed and the results of the classification of EEG signals shows the good performance compared with previous approaches.