Artificial Intelligence based prediction of seizures for Epileptic Patients: IoT based Cost effective Solution

Epilepsy prediction in epileptic patients are considered as the most crucial issue since decades. According to the world health organization (WHO) around fifty million people are suffering from epilepsy which is caused by the neurological disorder. Sudden and abrupt seizures are acknowledged as epilepsy. Seventy percent of the epileptic patients belong to the lower and middle-class countries. Seizures are very complex to predict as they are very sudden and it is very hard to discriminate the neurological signals. Seizures are due to the redundant discharge of electrical signals from brain cells. Several artificial intelligence based and deep learning based methodologies have been applied to detect seizures as they are dangerous to the growth of the brain. Classification and patterns of brain signals were classified using electroencephalogram (EEG) signals. Magnetic resonance imaging (MRI) of brain has also been performed to identify the flash floods. The identification of seizures in hospitals is acknowledged as expensive as the EEG signals and MRI testing equipment are very costly. According to the WHO mostly the epileptic patients belong to the poor countries therefore a cost-effective solution is needed. Seventy percent of the epileptic patients can live a normal life if it was properly diagnosed and treated. In this research paper a novel cost-effective solution has been presented for the vigorous prediction of the seizures in epileptic patients. Several methodologies and case studies have been studied by researchers but cannot be acknowledged as cost-effective as researchers used EEG and MRI signals. In this research paper a cost-effective hardware model has been designed and Modified cuckoo search has been applied to predict the seizures. Results have proved that the proposed research can be acknowledged as cost effective solution with higher accuracy and less false alarm rate.