Feature Extraction And Classification Of Eeg Signals Using Neural Network

Abstract

The use of Electroencephalogram (EEG) or “brain waves” for human-computer interaction is a new and challenging field that has gained momentum in the past few years. In this work different finite impulse response filter (FIR) windowing techniques (Rectangular, Hamming, Hanning, Blackman, Kaiser β= 5,8,12) are used to extract EEG signal to its basic components (Delta wave, Theta wave, Alpha wave, Gamma and Beta wave).The comparison between these windowing methods are done by computing the Fourier transform, power spectrum, SNR values. The features are extracted from the data and applied to classification techniques to identify the accuracy in obtaining the information of the data. In this research, EEG from one subject who performed four tasks has been classified using Radial Basis Function (RBF) and Multi Layer Perceptron (MLP) neural networks. Five data sets with 1000 samples are chosen in order to perform classification techniques. 200 iterations are done to identify the best error rate. These iterations help us to achieve best output. We calculate the elapsed time, confusion matrix, sensitivity, precision, specificity and accuracy for the classified data. The best classification accuracy is approximately 99.66% using the Multi Layer Perceptron technique and the best windowing technique obtained is Kaiser β= 12. The experimental results are performed using MATLAB Tool.

Authors and Affiliations

choppalli Sruth, prof. P. Mallikarjuna Rao

Keywords

Related Articles

Basis for an effective climate change adaptation: Case small Mexican agriculture

Mexico is considered one of the world's most vulnerable countries to climate change effects. Its geographical location and climatic, orographic and hydrological conditions contribute considerably to this situation. The o...

Digital Image Processing Assessment in Multi Slice CT Angiogram using Liner, Non-Liner and Both Classifiers

Nowadays with heart diseases most of peoples are dying lock process to find problem agile fashion in remote areas. In this research help to peoples who are staying remote also to find problem in heart on which location a...

An Innovative Technique to Utilize E-Plastic as Construction Material: A Review

Rapid growth of technology, up gradation in latest innovation of electrical industry in 21st century has led one of fastest growing waste stream in world commonly known as E waste. People became so techno geek that they...

The Influences of Structural Feature and Surface Properties of Montmorillonite on the Supported Olefin Polymerization Catalysts

Clay minerals possess unique physical properties such as high Lewis acidity, high specific surface area, periodic layer structures, and these materials can be utilized as polyolefin catalysts supports. This work focused...

Hybridizing Regular Expression with query processing to remove SQL Injection and XSS attacks

Web applications are widely used everywhere like e-commerce, online payments, online banking, money transfer, social networking, etc. As web application interacts with database where critical information is stored over t...

Download PDF file
  • EP ID EP416594
  • DOI 10.9790/9622-0811025360.
  • Views 165
  • Downloads 0

How To Cite

choppalli Sruth, prof. P. Mallikarjuna Rao (2018). Feature Extraction And Classification Of Eeg Signals Using Neural Network. International Journal of engineering Research and Applications, 8(11), 53-60. https://www.europub.co.uk/articles/-A-416594