EEG CLASSIFIACTION USING MACHINE LEARNING TECHNIQUES (SVM AND ELM)

Abstract

The EEG (Electroencephalogram) signal shows the electrical activity of the brain. They contain useful information about the brain state and there nature is highly random.

Electroencephalography (EEG) is method to record the electrical activity of the brain. It is typically non-invasive, with the electrodes placed along the scalp. EEG refers to the recording of the brain's spontaneous electrical activity over a period of time, as recorded from multiple electrodes placed on the scalp. Spectral content of EEG is generally focused by diagnostic applications.

In this thesis project, the two EEG signals that were analyzed were obtained by recording the EEG activity that was occurring when a person was moving their arms for the first case and their legs for the second case. A power line rejection notch filter, power spectral density analysis, Principal Component Analysis (PCA) were used to pre-processed these EEG signals, and finally, the mathematically based machine learning analysis is done using both Support Vector Machine (SVM) and Extreme Learning Machine (ELM). Machine learning is used to generate a model that could be used to accurately predict whether a person is moving their arms or their legs by applying the EEGs as inputs to the generated model and reading the output of the model.

The goal of this thesis is to compare the performance of SVM and ELM by using the accuracy of classification that each model produces.

Matlab results compared ELM and SVMs with classification accuracy of both models.