A NLU-based method for a first level automatic categorization of AI-based security documents

Abstract

Managing and organizing the information is a difficult subject. Especially in our time with the plethora of available sources accessible to the public, basically through the Internet, the volume of these pieces of information becomes larger and larger. The purpose of this dissertation is the creation of a model of documents' categorization. Our goal is to illustrate an as possible automated way of classification that takes into consideration the meaning of the documents. The way in which a human evaluates and classifies documents is followed in this proposed methodology. We also illustrate some empirical examples of the function of our classification method.