Identification of New Epilepsy Syndromes using Machine Learning

Abstract

Ubiquity of machine learning is a reality in the current world: machine learning is everywhere. Neurology is no exception. This paper presents an application of machine learning algorithms for the analysis of multi-national epilepsy clinical data. The initial purpose of the analysis was to find patters in the data, however the analysis resulted in the identification of two new epilepsy syndromes: Borderline Absence Syndrome and Childhood Myoclonic Epilepsy with Absence. It was confirmed that decision tree is an appropriate tool to present the results of supervised machine learning, helping the physicians make sense of the model and trace it back to the data.