Used various machine learning and signal processing methods to filter raw electroencephalogram (EEG) data and aid in the classification of epilepsy. I also tried to identify subtle brain connectivity patterns that can serve as signature patterns for epilepsy. Moreover, based on demography like age and sex, how these brain connectivity patterns changes were also proposed.Further, I tried to build a classification model that would, with about 93% accuracy distinguish intermediate EEG recordings from normal and epilepsy. Here, intermediate recordings refer to those where a physician is not able to correctly determine if the observed individual is epileptic or normal.In addition, also developed a machine learning based approach that was able to classify the above identified intermediate recordings as either epilepsy or normal with an F1 score of 87.3%.

Applied different rules and pattern mining techniques to identify human brain connectivity signatures that can help distinguish between epilepsy from normal electroencephalogram (EEG).