B.N. Onykij – Dr.Sc. (Eng.), Professor, Head of Department,
Department of Competitive Systems Analysis No. 65,
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
E-mail: bnonykij@mephi.ru
A.A. Artamonov – Ph.D. (Eng.), Head of Department of Competitive Systems Analysis No. 65,
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
E-mail: aartamonov@kaf65.ru
E.S. Tretyakov – Assistant,
Department of Competitive Systems Analysis No. 65,
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
E-mail: etretyakov@kaf65.ru
A.I. Cherkasskiy – Assistant,
Department of Competitive Systems Analysis No. 65,
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
E-mail: acherkassliy@kaf65.ru
K.V. Ionkina – Post-graduate Student, Department of Competitive Systems Analysis No. 65,
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
E-mail: kionkina@kaf65.ru

When conducting information and analytical studies in social networks tasks of searching for similar objects arise. Solving these tasks with search agent technologies is associated with the following problems: search engine of social networks is configured for interactive communication with a user, but not with his agent, since a priori the analytical tasks of the user are not known as well as the intellectual level of the search agent created by him; objects of social networks (persons and groups) are described with dozens of characteristics, moreover, some of them are optional, other filled characteristics may not be related to the target search, and finally, the characteristics of objects in social networks are described with all possible data types (numbers, dimension values, scores of qualitative characteristics, symbols, images, audio and video characteristics, texts in various national languages). Target agent search in such a complex information environment appears to be non-trivial task that does not have a common solution. To overcome the above problems, the authors propose inductive empirical models of target objects, that make it possible to conduct information and analytical studies in social networks with agent technologies correctly.

Objective of the study is the development and description of inductive (empirical) models of targets for training search agents of autonomous data collection in social networks.

The result is the proposed model for training agents in social networks is proposed. The agents are planned to solve problems of constructing target vectors and identify objects for various requests.

The solution of the problems described in the article allowed the authors to create a multi-agent multilingual information analytical system “Search” for conducting social research on youth subject using data from VK social network. The system was created on behalf of the Ministry of Science and Higher Education of the Russian Federation.

Onykij B.N., Artamonov A.A., Tretyakov E.S., Cherkasskiy A.I. Ionkina K.V. Inductive training models of search agents in social networks. Highly Available Systems. 2020. V. 16. № 1. P. 5–13. DOI: 10.18127/j20729472-202001-01.

1. Makarov V.L., Bahtizin A.R., Sushko E.D., Vasenin V.A, Borisov V.A., Roganov V.A. Agent-orientirovannye modeli: mirovoj opyt i tekhnicheskie vozmozhnosti realizacii na superkomp'yuterah. Vestnik Rossijskoj akademii nauk. 2016. T. 86. № 3. S. 252–262.
2. Kulik S. D., Protopopova J. Genetic Algorithm and Software Tools for Solving Optimization Problems in Intelligent Robotics. Advanced Technologies in Robotics and Intelligent Systems. 2020. P. 171–178.
3. Kulik S.D., Shtanko A.N. Experiments with Neural Net Object Detection System YOLO on Small Training Datasets for Intelligent Robotics. Advanced Technologies in Robotics and Intelligent Systems. 2020. P. 157–162.
4. Artamonov A.A., Kirichenko A.V., Tretyakov E.S., Ionkina K.V., Lopatina E.O., Cherkasskiy A.I. Agent-based search in social networks. International Journal of Civil Engineering and Technology. 2018. V. 9(13). P. 28–35.
5. Onykij B.N., Sokolina K.A. Konceptual'nye voprosy proektirovaniya mul'tiagentnyh informacionno-analiticheskih sistem dlya poiska i obrabotki nauchno-tekhnicheskoj informacii. Sistemy vysokoj dostupnosti. 2017. T. 13. № 1. S. 40–51.
6. Artamonov A.A., Leonov D.V., Onykij B.N., Pronicheva L.V. Mul'tiagentnaya informacionno-analiticheskaya sistema po estestvenno-nauchnym i tekhnologicheskim napravleniyam. Sistemy vysokoj dostupnosti. 2014. T. 10. № 2.