<![CDATA[Epilepsy is a chronic neurological disorder characterized by recurrent seizures caused by abnormal electrical activity in the brain. Fast and accurate seizure detection is crucial to support medical intervention and improve patients' quality of life. Currently, Electroencephalogram (EEG) signals are widely used to diagnose epilepsy as they record brain electrical activity in real-time. However, manual analysis of EEG signals requires time and precision, necessitating a more effective automated solution. This study aims to optimize the Backpropagation model for detecting epileptic seizures using EEG data. The research involved collaboration between Telkom University, Sumber Waras Hospital, and the University of Bonn. The EEG data collected was processed through Discrete Cosine Transform (DCT) to extract important features before being used to train the artificial neural network (ANN) model. The model was trained and tested using varying numbers of epochs to measure its accuracy. The results show that the Backpropagation model achieved optimal accuracy of 91.15% at 100 epochs and increased to 93.05% at 200 epochs. Although accuracy improved with more epochs, the longer computational time posed a risk of overfitting. This research demonstrates that the Backpropagation algorithm can be optimized to detect epileptic seizures accurately and efficiently. The implication for Sumber Waras Hospital is that this model can be implemented in EEG monitoring systems to detect seizures in real-time, supporting faster medical intervention and reducing reliance on manual analysis. Thus, this study contributes to providing a more efficient diagnostic solution and enhancing healthcare services for epilepsy patients.]]>
Copyrights © 2024
