Due to the fact that most seizures in persons with epilepsy occur seldom, it is critical for the classification and diagnosis of epilepsy that the EEG be used. When it comes to adult patients, empirical interpretation of a first EEG is quite insensitive, with sensitivities between 29 and 55% [29–55 percent]. During seizure-free EEG epochs, useful EEG data is buried within the signals, out of reach of any specialised physician in this area who is trained in EEG analysis. We develop a multi-variate strategy to understand the functional connectivity of the brain at the sensor level using EEG data in order to identify individuals with generalized epilepsy, in contrast to the majority of previous studies. Eight different connection characteristics were examined using five different measures across the temporal, periodic, and time-frequency domains. After evaluating the solution using the K-Nearest Neighbour approach, the results were compared to an epilepsy group, and subsequently to a group of patients who had non-epileptic episodes. Classification accuracy (89%) was achieved for EG and HC, however substantial spatial-temporal deficits in the front central areas in the beta frequency band were found in the EG group compared to the HC group. Because of the well-documented coexistence of NEAD and epileptic episodes, the classification accuracy for EG and NEAD was only around 79%. People with specialised epilepsy may be consistently distinguished from those with HC using seizure-free EEG data, according to this research. Although additional study is required in this area to establish a diagnostic tool that is therapeutically helpful,

Epilepsy, EEG, Brain Signals, Nero Machine Learning, Health care, K-Nearest Neighbour