An Innovative Method for Brain State Classification from EEG Data using Hybrid Learning Algorithm

Volume 3 Issue 4 December - February 2017

Research Paper

An Innovative Method for Brain State Classification from EEG Data using Hybrid Learning Algorithm

Nithiya*, 0**, E. Manoj***

* Department of Embedded System Technologies, K.S.R. College of Engineering, (Autonomous), Tamil Nadu, India.

** Lecturer, Department of Electrical and Electronics Engineering, C.M.S. Group of Institutions, Tamil Nadu, India.

*** UG Scholar, Department of Electrical and Electronics Engineering, Coimbatore Institute of Technology (Autonomous), Tamil Nadu, India.

Nithya., Sabarinathan, E., and Manoj, E. (2017). An Innovative Method for Brain State Classification from EEG Data using Hybrid Learning Algorithm. i-manager’s Journal on Pattern Recognition, 3(4), 8-15. https://doi.org/10.26634/jpr.3.4.13538

Abstract

In biomedical signal analysis, classification plays an important role and gives the promising solution to the electroencephalogram (EEG) analysis. An automatic EEG signal classification is proposed in this system and contribution of the diagnostician is replaced by using the soft computing techniques, since the manual classifications carried out in the clinical analysis is a time consuming task. In the proposed methodology, the EEG features are extracted from the raw EEG signals which are then fed to these ANFIS classifiers. It is a sophisticated framework for the classification of the different brain states in the human brain by representing their experts based knowledge as an Adaptive Neuro Fuzzy Inference System (ANFIS). This algorithm has a capability to detect the two types of brain states, including dementia and encephalopathy. Finally, the tentative outcome of the results is expressed in terms of classification accuracy and improves execution. The analyses are demonstrated with the ANFIS algorithm to improve and enhance the performances in the MATLAB.