Power spectral density and KNN based adaboost classifier for epilepsy classification from EEG

In the arena of biomedical engineering, the classification and analysis of epilepsy from Electroencephalography (EEG) signals forms an important area of research. When the neurons get hyper excited, seizures occur causing a lot of inconvenience and trouble to the patient. For the study of the predominant abnormalities in the cerebral cortex of the brain, EEG is used widely. Due to the long nature of the EEG recordings, it is very difficult for the clinicians and visual experts to analyze the entire waveforms and hence automated detection of seizures from EEG signals came into existence. In this paper, the dimensions of the recorded EEG signals was reduced with the help of Power Spectral Density (PSD) and then the dimensionally reduced values are classified with the help of KNN Based Adaboost Classifier for the classification of epilepsy and the performance metrics are analyzed. Results show that an average accuracy of about 97.53% along with an average Performance Index of about 94.85% is obtained.