Method of Identification of Extremist Texts in the Russian Language Based on the Fuzzy Logic

摘要

The article describes the author's methodology of the study of the Russian language texts in order to identify their extremist orientation. This method is based on the fuzzy logic. The identified problems in solving the task are indicated. Then the description of the method is presented. There are also main points of fuzzy logic and soft computing tools used in the development of the method. We also determined a set of indicators consisting of four groups: (1) markers of extremist orientation (by types of extremism); (2) markers of linguistic manipulation; (3) indicators of verbal aggression; (4) scientific and journalistic terms. There is quantitative and qualitative characterization for each group of indicators. After that, the process of formation of the evaluation of the studied text on the extremist orientation is described in detail. This process is described on the basis of the system being fuzzy-logical conclusions, called fuzzy multilevel [0,1] - classifiers.