A.V. Proletarsky - Dr.Sc. (Eng.), Professor, Dean of Faculty «Informatics and control systems», Bauman Moscow State Technical University E-mail:pav@bmstu.ru M.A. Skvortsova - Research Assistant, Computer Systems and Networks Department, Bauman Moscow State Technical University E-mail: magavrilova@bmstu.ru V.I. Terekhov - Ph.D. (Eng.), Associate Professor, Information Processing and Control Systems Department, Bauman Moscow State Technical University E-mail: terekchow@bmstu.ru

The purpose of this article is to analyze approaches to hybrid intelligent risk assessment system design based on unstructured information. Hybrid intelligent systems are able to analyze all of the risks existing in the sphere analyzed, form an evaluation report on the risks found and make recommendations on protection measures. As study area, the author selected the field of natural and technological hazards. To conduct research in the field of risk assessment in natural and technogenic spheres, the basic concepts used in this field were introduced. As input data for the analysis of information, the data from 40 sites and more than 90 sections on those sites were used (federal and regional sites with open data, information agencies and news portals, websites with analytical information). The article presents a generalized structure of the hybrid intelligent system of risk assessment based on heterogeneous information. Practical experiments for the risk assessment were conducted on the data obtained from unstructured sources using the following techniques: risk assessment method on the event-based methodology, risk assessment based on indicators, and general risk assessment of natural and technogenic spheres. The methodology of risk assessment based on the events is based on the method of graphs. The methodology of risk assessment based on indicators takes as basis the calculation of the average interval of the average for each area of life activity, the assignment of weights for indicators in the range [0,1] and setting the measure values in the range [0,1]. The analytic hierarchy process and the method of paired comparison form the basis of an overall risk assessment. Special attention is paid to the assessment of relative criteria importance. The major challenge of the analysis of hybrid intelligent systems design is related to the evaluation of risk and its factors (threats, hazards, vulnerabilities) and caused by the following problems: 1) Incomplete information about the components of risk and their ambiguous properties; 2) The complexity of creating a model of information system and evaluation of its vulnerability; 3) The duration of the evaluation process and rapid loss of relevance of its results; 4) The complexity of data aggregation from various sources, including statistical information and expert assessments; 5) The need for the involvement of several specialists on risk analysis to improve the adequacy of assessments. In conclusion, the author states that the design of the hybrid intelligent risk assessment system based on unstructured information is a difficult and time-consuming engineering task. Therefore, to obtain better results, it is necessary to rely on a number of criteria, such as the use of more techniques, creating powerful expert system, application of other intelligent methods.

 

1. Polozhenie o sisteme nezavisimojj ocenki riskov v oblasti pozharnojj bezopasnosti, grazhdanskojj oborony i zashhity naselenija i territorijj ot chrezvychajjnykh situacijj prirodnogo i tekhnogennogo kharaktera na territorii rossijjskojj federacii. Rezhim dostupa: http://41.mchs.gov.ru/document/2912473 (data obrashhenija 05.12.2016)
2. CHernenkijj V.M. Algoritmicheskaja model opisanija diskretnogo processa funkcionirovanija sistemy // technomag.edu.ru: Nauka i Obrazovanie: ehlektronnoe nauchno-tekhnicheskoe izdanie. 2011. Vyp. 12. URL http://technomag.edu.ru/doc/292997.html.
3. Xiao X., Zhang H., Hasegawa O. Density Estimation Method Based on Self-Organizing Incremental Neural Network and Error Estimation // Proceedings of the Neural Information Processing: 20th International Conference, ICONIP 2013. Daegu, Korea. 2013. R. 43-50.
4. Kolesnikov A.V., Kirikov I.A., Listopad S.V. Gibridnye intellektualnye sistemy s samoorganizaciejj: koordinacija, soglasovannost, spor. M.: IPI RAN. 2014. 189 s.
5. GOST R 51898-2002. Aspekty bezopasnosti. Pravila vkljuchenija v standarty: sochetanie verojatnosti nanesenija ushherba i tjazhesti ehtogo ushherba.
6. The IEEE website. [Online]. Available: http://www.ieee.org/
7. CHechkin A.V. Intellektualnaja informacionnaja sistema na osnove radikalnogo modelirovanija kak instrumentalnoe sredstvo obespechenija kompleksnogo razvitija // Nejjrokompjutery: razrabotka, primenenie. 2015. № 5. S. 7-13.
8. Tarasov V.B. Ot mnogoagentnykh sistem k intellektualnym organizacijam: filosofija, psikhologija, informatika. M.: EHditorial URSS. 2002. 352 s.
9. Pugh J.K., Stanley K.O. Evolving Multimodal Controllers with HyperNEAT. Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2013). New York, NY: ACM, 2013. 8 p.
10. Xu SW-J., Dong Y-C., W-L. Xiao. Is It Reasonable for Saaty\'s Consistency Test in the Pairwise Comparison Method - // Computing, Communication, Control, and Management, 2008. CCCM \'08. 2008 ISECS International Colloquium on Computing, Communication, Control, and Management. Aug, 2008. V. 3. P. 294-298, 3-4.
11. Riza S., Murtuzayeva M. Application saaty pair comparisons method to the investments distribution in parameters of ecological sustainability // Problems of Cybernetics and Informatics (PCI). 2012 IV International Conference «Problems of Cybernetics and Informatics» (PCI). Sept. 2012. P. 1-3.