Classification Of Eeg Signal Data Using Hybrid Of Neuro-Fuzzy
Pratibha Rana · Zenodo (CERN European Organization for Nuclear Research) · 2016
Brain Computing interface technology represents a very highly growing field now-a-days for the research because of its unique applications system. In this paper we investigate classification methods of mental commands based on EEG data for BCI. The aim of this study is to present the work of training an artificial neural network (ANN) and Neuro-Fuzzy system provided with the data of a few healthy people who are in different mental states thinking about different activities like eating, walking or sleeping etc. It creates a mutual understanding between mind wave signals and the machine or surrounding system. In recent years’ applications based on the brain computing interface have gained a rather huge popularity because of its benefits like providing disabled people with a communication and control environment along with different types of movement restorations. It is proving to be revolutionary in the fields of medical and robotics, mind reading and remote communication etc. This paper puts forward the idea of using a classification algorithm based on a new hybrid approach using neural networks and fuzzy logic that can collectively detect and recognize the state of mind. The EEG brain signals tend to change as our mind state changes because of the voltage fluctuations resulting from the ionic currents in the neurons of the Brain. Features are extracted from the raw electroencephalography (EEG) data using data processing technique PPCA (probabilistic principle component analysis) and are fed to the classifier. The features extracted have been used to train the neural network by using the MATLAB toolbox (Bronzino, J. D. 2010). The work carried out here is to be later on compared with the outputs generated by using just a neural networks based classification algorithm.