A Machine learning approach to monitor Brain Health and epilepsy detection

Abstract

Brain health is a major concern worldwide and in India. There are a ton of neurological issues that may influence the mind and one such issue is epilepsy. Epilepsy is one of the preeminent generally happening neurological issues. Electroencephalography (EEG) signals help in the examination and conclusion of epilepsy by recording the exercises inside the cortical districts of the epileptic patient. Conventional techniques for breaking down an EEG signal for epileptic seizure discovery are tedious. A proposed work of robotized seizure recognition systems utilizing an AI strategy is proposed to supplant these conventional techniques. The proposed work is completed with a real-time dataset of epileptic patients taken at a different period. The two basic steps involved during this project are pre-processing and classification. Information pre-preparing is an information mining procedure that includes changing crude information into a reasonable arrangement. Classification is the process of predicting the class of a given dataset. The classifiers used are SVM, Naive Bayes, Random Forest classifier. The random forest is most preferred for its decision trees where the model has low variance and also for its high predictive performance. The proposed system helps in daily monitoring the brain health of epileptic patients using a mobile application. This system mainly focuses on rehabilitated patients. Once when the patients use a wireless wearable headband the signal is monitored and their health condition is detected using a mobile application.