Analysis on Wavelet Feature and Softmax Discriminant Classifier for the detection of epilepsy. Issue 1 (March 2021)

• Record Type: Journal Article
• Authors: Deepa, R; Rajaguru, Harikumar; Ganesh Babu, C
• Is Part Of: IOP conference series. Volume 1084:Issue 1(2021)
• Publication Date: 2021-03
• Subjects: EEG Signals -- Haar -- dB2 -- Sym8 -- dB4 -- Softmax Discriminant Classifier; Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/1084/1/012036
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: The most frequently diagnosed brain disease is epilepsy, which is characterised by the unexpected onset of frequent seizures. The detection of epilepsy in this paper was established by using the wavelet features Haar, dB2, Symlets (Sym8) and dB4, followed by the Softmax Discriminant Classifier, which uses to detect the epilepsy from the EEG signals. The performance of the wavelet features and classifier is evaluated based on the performance index, specificity, sensitivity, precision, time delay and quality values. Amongthe wavelet features, the sym8 performs better than the other and processed further using the Softmax Discriminant Classifier, which outperforms the 90.93 percent classification accuracy, with a low time delay of 1.991s, the 72.61 percent output index, the most promising result in this work.