Epileptic seizure detection using deep learning through min max scaler normalization

Epileptic seizure detection and prediction are significantly sought-after research currently because robust algorithms are available. Machine learning and deep learning have allowed us to analyze brain signals with high accuracy. The brain signals collected using EEG (electroencephalogram) are complex and prone to noise. This paper describes a pre-processed dataset created using the famous CHB-MIT scalp EEG database. A deep learning model is trained and tested by applying the Bidirectional Long Short Term Memory (BiLSTM) algorithm through MinMaxScaler normalization on this pre-processed dataset. The results from this published dataset and model are promising in terms of accuracy, precision, and F1 score when compared with earlier research works. Accuracy is 99.55%, precision is 99.64%, and F1 score is 99.52% for the proposed model when the seizure activity data is considered for all the patients.