Enhancement of EEG Signal Feature Extraction and Classification Using Machine Learning Techniques

Abstract

EEG signal analysis is one of the active research areas in medical analysis and diagnosis. EEG analysis is used to diagnose various diseases, to find out the psychological status of people, feeling state of people, and also used to design human-computer interfaces to control computers through thoughts. There are many research challenges in this EEG signal analysis. An efficient pre-processing technique is required to filter out channel noise that is picked during acquisition. This is one of the critical tasks in EEG signal analysis because the presence of noise will affect the accuracy of the Analytic result. newlineThe accuracy of signal analysis also depends upon the feature vector involved in the signal analysis, especially when we are employing a machine learning based classifier for this purpose. Moreover, one type of feature vector will give better performance for a kind of application and will not perform well if we change the application. If we can find a new feature vector that will be good for the many application, then it will be more useful. Understanding these factors into account, a research hypothesis is formulated such that will it be possible to generate a new feature vector that will outperform diverse applications newline