Intelligent system for evaluating destructive nature of text content of social networks based on fuzzy logic

Abstract

The article deals with the actual problem of counteraction and timely detection of destructive information, which has become widely distributed on the Internet, which, in connection with the significant expansion of Internet communication facilities, poses a threat to society as a whole. In this paper we consider the problem of constructing an intellectual system for evaluating the destructive influence of text content of social networks under uncertainty on the basis of application of product rules of fuzzy logic. An approach to building an intellectual system based on the neural network and functions of the fuzzy decision making system is described, taking into account estimates of linguistic analysis of the text as input data of the system. Among the methods of linguistic analysis of the text are two main ones used in the evaluation of text content of social networks, such as: determining the emotional color of the text and the classification of the text. The main terms of the linguistic variables with the specified values ranges are to be taken into account for assessing the destructive impact of the text content of social networks. The main steps of building an intellectual system based on a combination of a neural network and a fuzzy logic device are identified, which in the end represents a hybrid network. The structure of the hybrid network is described. The method of teaching a neural network based on the rules of fuzzy logic is given. The stages of the transformation of fuzzy data in the process of logical decision-making are specified, and their implementation in the Matlab environment is demonstrated. An example of an adequacy check of the proposed hybrid network is given.