A different approach to epilepsy risk level classification utilizing various distance measures as post classifiers

Abstract

The electrical activity of the brain can be studied thoroughly through the recordings of the Electroencephalography (EEG) signals and is considered as a vital tool for the analysis and diagnosis of neurological diseases like tumours of the brain, epilepsy and other cognitive disorders. Due to the continuous electrical discharges from the cortex of the cerebrum, epilepsy occurs which results in several severe consequences thereby making many vital changes in the EEG signal. In this paper, the epilepsy risk levels are classified by making use of Approximate Entropy as a Feature Extraction technique followed by Various Distance Measures such as Euclidean Distance Measure (EDM), City Block Distance Measure (CBDM) and Correlation Distance Measure (CDM) as Post Classifiers for the perfect classification of epilepsy risk levels from EEG signals. The validation parameters taken here are Performance Index (PI), Time Delay (TD), Quality Value (QV), Sensitivity, Specificity and Accuracy.