Social networks have become a part of our modern life. They are easy to use and versatile, which helps to attract more and more people every day. However, along with the obvious benefits social networks pose a number of problems. One of these problems is the open access publication of provocative and anti-social posts used to influence the public. Often such posts contain calls for ethnic discord, anti-social activities, promotion of violence, and radical religious views. As a rule, the moderation of posts in social networks is reduced to the search for spam, flooding, off-topic posts and comments, usually containing advertising. Due to the exponential growth of the data posted by users the automated moderation process has become an urgent problem. Both foreign and domestic developers are involved in active research to solve this problem. The authors suggest using the technology of multi-agent systems (MAS) and software agents to build a system for monitoring users' posts on the Internet in order to identify security threats. At the same time, an independent task is to develop thematic ontologies containing a description of threats in the selected domain. This article describes a conceptual model of multi-agent system for users’ posts moderation and a structure of MAS ontology used as a basis for a fragment of ontology describing the type and direction of security threats. The ontology is a description of a partially ordered set of concepts to be used by agents detecting security threats. Each agent specialty is reflected through a subset of concepts, some of which may be shared by several agents. The described fragment of ontology allows agents to use one and same terms in the communication process and clearly perceive working terminology. The authors have chosen Protégé ontology editor as a tool for developing the ontology and OWL language as a modeling technique. It is shown that the use of agent-based approach greatly simplifies software design, because it enables working with data as knowledge, taking into account the context through the use of ontologies.
ontologii, agenty, mnogoagentnye sistemy
1. Analiz ontologii vzaimodeystviya MAS pri reshenii zadach obespecheniya informacionnoy bezopasnosti. URL: http://progrm.ru/?p=260
2. Antispam prilozheniya dlya Web-saytov ot CleanTalk. URL: http://cleantalk.ru
3. Boldyrev E.V., Kiryakova G.S., Shilkin A.V. Agentnyy podhod k proektirovaniyu setevyh prilozheniy poiska informacii // Vychislitel'nye tehnologii. T. 10. Special'nyy vypusk. 2005.
4. Bolotova L.S. Sistemy iskusstvennogo intellekta: modeli i tehnologii, osnovannye na znaniyah: Uchebnik. M., 2012.
5. Gribova V.V., Kleschev A.S., Krylov D.A., Moskalenko F.M., Timchenko V.A., Shalfeeva E.A. Agentnyy podhod k razrabotke intellektual'nyh Internet-servisov // Trudy Kongressa po intellektual'nym sistemam i informacionnym tehnologiyam «IS&IT’12». M., 2012. T. 1.
6. Nikitin A.G. Vidy i klassifikacii ekstremistskoy deyatel'nosti: nekotorye pravovye aspekty // Tatischevskie chteniya: aktual'nye problemy nauki i praktiki: Materialy VI Mezhdunarodnoy nauchno-prakticheskoy konferencii (16-19 aprelya 2009 g.). Ch. 1. Tol'yatti, 2009.
7. Ohapkina E.P., Lukoyanov I.A., Voronov V.I., Voronova L.I. Razrabotka i vnedrenie poiskovogo robota dlya analiza interesov klientov // Studencheskiy nauchnyy forum: VI Mezhdunarodnaya studencheskaya elektronnaya nauchnaya konferenciya (15 fevralya - 31 marta 2014 g.). URL: http://www.scienceforum.ru/2014/495/4758
8. Tanenbaum E., Van Steen M. Raspredelennye sistemy. Principy i paradigmy. SPb., 2003.
9. Comment E-Mail Verification. URL: http://wordpress.org/extend/plugins/comment-email-verify
10. Okhapkina E.P, Voronova L.I. The development of the ontology for a multi-agent subsystem analyzing user posts in social networks in order to identify security threats to society // International Journal of Applied and Fundamental Research. 2013. № 2. URL: http://www.science-sd.com/455-24393
