Heterogeneous Semi-Structured Data Analysis in Information Security

Abstract

When solving information security problems, one needs to analyze and evaluate the similarity of heterogeneous semi-structured objects sets. The authors propose to use for this an approach on the basis of case analysis. The data is represented as a set of objects (use cases), each of which is described by a variable set of properties. The paper proposes measures to evaluate the similarity of individual objects and precedents, presents the results of the similarity evaluation method experimental testing. The authors present the architecture of the analysis system and the method of its learning.