• C Nadhamuni Reddy
• E D Pavankumar
• T Chaitanya
• A Bindu Priya
• M Chamanthi,D Lakshman Ritesh Babu

These are called recurrent seizures and are caused by brain's electrical activity and the condition is termed as epilepsy which is a very common neurological disorder. Adequate prediction and early detection of seizures on the other hand, play a vital role in effective intervention, thus improving patient outcomes and quality of life. It can be used to monitor the brain activity and diagnosis of epilepsy through using electroencephalogram (EEG) signals. Prediction of seizure has traditionally been based on use of traditional machine learning methods such as Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), and Random Forests. While methods of this type may not require manual feature extraction, they are domain expertise dependent and cannot be generalized to datasets separated in more ways from each other than being similar in that they are collections of videos or audio tracks. In addition, the usual models cannot characterize the complex temporal relations in EEG signals, leading to the lower prediction accuracy and reliability in the real–time application.This work presents a deep learning method based on Recurrent Neural Networks (RNNs) and Long Short Term Memory (LSTM) to deal with issues of epileptic seizure prediction. Especially, these models are particularly designed to deal with sequential data that is more suitable for the analysis of EEG signals. State of the art techniques of feature extraction are used to preprocess EEG data for the designed system in order to emphasize the patterns that are of more value in detecting a seizure. Long term dependencies in brain activity will be modeled by LSTM and RNN networks, and in some cases can leverage some of the features of the EEG signals to be used to detect some of the subtle changes that precede precursory seizures as early warning alerts for patients and caregivers.In addition, the proposed deep learning framework has many benefits over existing systems. It learns complex patterns directly from the raw EEG signal thus improving the prediction accuracy and there is no need for manual feature extraction. It has been optimized for real time for the purpose of seizure detection in these clinical and homebased settings. Finally, this EEG dataset is trained on multiple datasets and in particular, becomes more generalized to work with other demographics and clinical coordination in practice. Additionally, the system can be integrated with deep learning techniques for scalability, and used also to extend the system with the larger datasets and the emerging EEG monitoring technologies.

Deep Learning Techniques, EEG Signals, Epileptic Seizures, Real-time Signal, Seizure Prediction, Epileptic Seizure Prediction, Real-time EEG Signals, Signal Processing,Temporal Features, Wavelet Transform, Domain Features, Seizure Onset, Frequency Domain Features, Epileptic Patients, EEG Signal Processing.

"EPILEPTIC SEIZURE DISORDER PREDICTION USING DEEP LEARNING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 4, page no.f374-f379, April-2025, Available :http://www.jetir.org/papers/JETIR2504549.pdf

"EPILEPTIC SEIZURE DISORDER PREDICTION USING DEEP LEARNING", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 4, page no. ppf374-f379, April-2025, Available at : http://www.jetir.org/papers/JETIR2504549.pdf