International Journal of Soft Computing
ISSN: OnlineISSN: Print 1816-9503
Abstract
Epilepsy can be diagnosed using technologies like Electroencephalogram (EEG), Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), etc. In this research, researchers inspect EEG signals for seizures. The seizure is recognized with the support of Independent Component Analysis and the ascertained signals are trained under supervision by making use of neural networks technique namely backpropagation algorithm. In the proposal, performance of network is evaluated using publicly available EEG dataset for various backpropagation training functions such as Gradient Descent Algorithm (GD), Scaled Conjugate Gradient (SCG), One Step Secant (OSS), Powell-Beale Restarts (PBR), Gradient Descent with Adaptive (GDWA), Fletcher-Powell Conjugate Gradient (FPCG) and Levenberg Marquardt (LM) Backpropagation are used here for the comparison technique. On comparing the performance of these aforementioned algorithms, highest accuracy with lowest mean square error was obtained for scaled conjugate gradient.
How to cite this article:
K. Sivasankari and K. Thanushkodi . Performance Comparison for MLP Networks Using Various Back Propagation Algorithms in Epileptic Seizure Detection.
DOI: https://doi.org/10.36478/ijscomp.2013.207.217
URL: https://www.makhillpublications.co/view-article/1816-9503/ijscomp.2013.207.217