MODULE FOR INTELLIGENT ANALYSIS OF UNSTRUCTURED TEXT DATA AS A SOUND TECHNICAL SOLUTION THAT EXPANDS THE FUNCTIONALITY OF DLP-SYSTEMS

Abstract

In this article, in accordance with the conducted research of the ways of functioning of the subsystem of content analysis of documents of information leakage prevention systems based on signature methods, its structural-functional and mathematical models are developed and presented. The conditions for achieving the maximum value of the classification quality index of this subsystem are formalized, its main disadvantages are highlighted, including those manifested in the implementation of insider destructive influence on protected information. The justification of the need to use intelligent methods in the process of content analysis is given. As a technical solution that expands the functionality of the subsystem of content analysis of documents based on signature methods, a module for intelligent analysis of unstructured text data is proposed, which allows binary classification of unstructured text data of minimal volume according to the degree of confidentiality in compliance with a given classification quality threshold. The requirements for the functioning of this module are verbally described and formalized. As part of the creation of the module, the task of two-stage optimization is formulated and presented, which consists in maximizing its efficiency function.