The combined method of textual information analysis for the content of destructive indicators

Abstract

Modern society is characterized by increased progress in science and technology, accompanied by rapid uncontrolled growth and a variety of textual information, including of destructive orientation, in the information space of the Internet. In this regard, as well as in the context of the potential creation of a “closed” Internet space in the Russian Federation, solving the problem of analyzing text content by searching and automatically identifying destructive information is a priority trend. Within the framework of the paper task of identifying destructive information is reduced to the task of classifying the analyzed textual data based on the presence or absence of destructive indicators. The article describes the universal three-stage combined method, including blocks for normalizing input data, the modified dictionary search of destructive indicators and the Bayesian classifier, which allows staged search of various indicators in the text and automatically classifies it as a “dangerous”.