Abnormalities analysis of EEG Signals for Seizure detection using Logistic regression model and a comparative analysis of neural nets and support vector machine.

  • Ruchi Sharma and Khyati Chopra

Abstract

 The brain is the first entity to detect the abnormalities by electroencephalogram to record the electrical pulses. The variations in EEG can easily detect the disorders of brain like seizures, brain tumours, epilepsy, Parkinson, alzeihmers etc. The doctors can easily detect these disorders by observing, EEG patterns of the patients. To detect the brain diseases is a challenging process to analyze. We propose a dynamic system to detect the brain abnormalities. Proposed work is a useful tool to analyse the normal and abnormal seizure patients. The logistic regression is a linear classier.it is a forward approach to classify different sets of data and their belong-ability. The accuracy of this system can be analysed easily while looking at the different eeg waves i.e. beta, alpha, theta, delta etc .This accurate detection of abnormality will reduce the efforts by doctors for diagnosing the disorder. The maximum accuracy attained by svm algorithm is 98.34% and 95.65 by the neural nets, considered to be a good accuracy rate. This suggested system obtained a better accuracy rate .

Published
2020-05-15
How to Cite
Ruchi Sharma and Khyati Chopra. (2020). Abnormalities analysis of EEG Signals for Seizure detection using Logistic regression model and a comparative analysis of neural nets and support vector machine. International Journal of Advanced Science and Technology, 29(12s), 1418 - 1429. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/22645
Section
Articles