Method of Complex Assessment of the Level of Information Content of Classification Features in the Conditions of Fuzzy Data Structure

The purpose of research is to develop a method for a comprehensive assessment of the level of information content of classification features in conditions of incomplete and fuzzy data structure.

Methods. The methodology of synthesis of hybrid fuzzy decision rules is used as a basic mathematical apparatus, focused on decision-making in conditions of incomplete and fuzzy description of the analyzed data. In the absence of training samples, it is proposed to use expert evaluation using brainstorming using the Delphi method and the theory of measuring latent variables with the G. Rush model to assess informativeness. In the presence of training samples, including small-volume samples - methods of expert evaluation, the informative Kullback measure, discriminant analysis, the method of group accounting of arguments, the Wald method, and the theory of measurement of latent variables.

Results. As the basic elements of mathematical models for calculating indicators of informativeness, it is proposed to use normalization functions of informativeness that take into account measures of confidence in the methods and data used and have the properties of membership functions. The desired indicators of informativeness for individual features and for the entire feature space are obtained by aggregation of normalization functions, and aggregation functions are selected taking into account the specifics of the tasks being solved.

Conclusion. In the course of the conducted research, the task of developing a method for a comprehensive assessment of the level of information content of classification features in conditions of incomplete and fuzzy data structure was solved. It was shown that to quantify the informativeness of each of the features, several methods describing various properties of the data structure under study should be used, and the integral indicator of informativeness should be determined by synthesizing fuzzy models in accordance with the general recommendations of the methodology for synthesizing hybrid fuzzy decision rules. To improve the accuracy of decision-making models, a mechanism for their correction is proposed by introducing confidence measures to the composition of informative features and the volume of the training sample.

1. Duke V., Emanue V. Informacionnye tekhnologii v mediko-biologicheskih issledovaniyah [Information technologies in biomedical research]. St. Petersburg, Peter Publ., 2003. 528 р.

2. Zhuravlev Yu. I., Gurevich I. B. Raspoznavanie obrazov i analiz izobrazhenij [Pattern recognition and image analysis]. Iskusstvennyi intellekt [Artificial intelligence]; ed. by D. A. Pospelov. Moscow, Radio and Communications, 1990, pp. 149-190.

3. Zagoruiko N. G. Prikladnye metody analiza dannyh i znanij [Applied methods of data and knowledge analysis]. Novosibirsk, Publishing House of the Institute of Mathematics, 1999. 270 p.

4. Kapustina S. V., Kiryakova O. V., Kapustina A. V., Lapina L. A., Stupina A. A. Vybor informativnyh priznakov dlya ocenki tyazhesti zabolevaniya [The choice of informative signs for assessing the severity of the disease]. Sovremennye problemy nauki i obrazovaniya = Modern Problems of Science and Education, 2015, по. 2-2, pp. 55.

5. Kolesnikova S. I. Metody analiza informativnosti raznotipnyh priznakov [Methods of analyzing the informativeness of different types of signs]. Vestnik Tomskogo gosudarstvennogo universiteta. Upravlenie, vychislitel'naya tekhnika i informatika. Obrabotka informacii = Bulletin of Tomsk State University. Management, Computer Engineering and Computer Science. Information Processing, 2009, no. 1(6), pp. 69-80.

6. Bunyaev V. V., Korenevsky N. A. Metody poiska informativnyh proekcionnyh zon i sinteza nechetkih reshayushchih pravil dlya refleksodiagnostiki [Methods of searching for informative projection zones and synthesis of fuzzy decision rules for reflexology]. Sistemnyj analiz i upravlenie v biomedicinskih sistemah = System Analysis and Management in Biomedical Systems, 2004, vol. 3, no. 3, pp. 175-178.

7. Vorontsov I. M., Shapovalov V. V., Sherstyuk Yu. M. Zdorov'e. Opyt razrabotki i obosnovanie primeneniya avtomatizirovannyh sistem dlya monitoringa i skriniruyushchej diagnostiki narushenij zdorov'ya [Health. Experience in the development and justification of the use of automated systems for monitoring and screening diagnostics of health disorders]. St. Petersburg, IPK Costa Publ., 2006. 432 р.

8. Korenevsky N. A., Rodionova S. N., Khripina I. I. Metodologiya sinteza gibridnyh nechetkih reshayushchih pravil dlya medicinskih intellektual'nyh sistem podderzhki prinyatiya reshenij [Methodology of synthesis of hybrid fuzzy decision rules for medical intelligent decision support systems]. Stary Oskol, TNT Publ., 2019. 472 р.

9. Korenevsky N. A., Rodionova S. N., Khripina I. I., Myasoedova M. A. Gibridnye nechetkie modeli ocenki funkcional'nogo sostoyaniya i sostoyaniya zdorov'ya cheloveka-operatora informacionno nasyshchennyh sistem [Hybrid fuzzy models for assessing the functional state and health of a human operator of information-saturated systems]. Sistemnyj analiz i upravlenie v biomedicinskih sistemah = System Analysis and Management in Biomedical Systems, 2019, vol. 18, no. 2, pp. 105-109.

10. Korenevsky N. A., Artemenko M. V., Provotorov V. Ya., Novikov L. A. Metod sinteza nechetkih reshayushchih pravil na osnove modelej sistemnyh vzaimosvyazej dlya resheniya zadach prognozirovaniya i diagnostiki zabolevanij [Method of synthesis of fuzzy decision rules based on models of system relationships for solving problems of forecasting and diagnosis of diseases]. Sistemnyj analiz i upravlenie v biomedicinskih sistemah = System Analysis and Management in Biomedical Systems, 2014, vol. 13, no. 4, pp. 881-886.

11. Korenevskiy N. A., Degtyarev S. V., Seregin S. P., Novikov A. V. Use of an Interactive Method for Classification in Problems of Medical Diagnosis. Biomedical Engineering, 2013, vol. 47, is. 4, pp. 169-172.

12. Korenevskiy N. A. Application of Fuzzy Logic for Decision-Making in Medical Expert Systems. Biomedical Engineering, 2015, vol. 49, pp. 46-49.

13. Korenevsky N. A., Rodionova S. N., Govorukhina T. N., Myasoedova M. A. Nechetkie modeli ocenki urovnya ergonomiki tekhnicheskih sistem i ee vliyanie na sostoyanie zdorov'ya cheloveka operatora s uchetom funkcional'nyh rezervov [Fuzzy models for assessing the level of ergonomics of technical systems and its impact on the human health of the operator, taking into account functional reserves]. Modelirovanie, optimizaciya i informacionnye tekhnologii = Modeling, Optimization and Information Technologies, 2019, vol. 7, no. 1 (24), pp. 39-53.

14. Korenevsky N. A., Bykov A. V., Tsymbal E. V., Aksenov V. V., Rodionov D. S. Prognozirovanie vozniknoveniya i recidiva insul'tov golovnogo mozga na osnove gibridnyh nechetkih modelej [Forecasting the occurrence and recurrence of brain strokes based on hybrid fuzzy models]. Modelirovanie, optimizaciya i informacionnye tekhnologii = Modeling, Optimization and Information Technologies, 2018, vol. 6, no. 3(22), pp. 50-72.

15. Korenevsky N. A., Shutkin A. N., Gorbatenko S. A., Serebrovsky V. I. Ocenka i up- ravlenie sostoyaniem zdorov'ya obuchayushchihsya na osnove gibridnyh intellektual'nyh tekhnologij [Assessment and management of students' health on the basis of hybrid intelligent technologies]. Stary Oskol, TNT Publ., 2016. 472 р.

16. Korenevsky N. A., Tutov N. D., Lazurina L. P. Proektirovanie mediko-ekologicheskih informacionnyh sistem [Design of medical and environmental information systems]. Kursk, Kursk State Technical University Publ., 2001. 193 p.

17. Korenevsky N. A., Ivankov Yu. A., Yakovleva E. A., Savchenko N. N. Sintez nechetkih reshayushchih pravil dlya prognozirovaniya i rannej diagnostiki zabolevanij, vyzyvaemyh sostoyaniem okruzhayushchej sredy, s uchetom individual'nyh osobennostej organizma [Synthesis of fuzzy decision rules for forecasting and early diagnosis of diseases caused by the state of the environment, taking into account individual characteristics of the organism]. Sistemnyj analiz i upravlenie v biomedicinskih sistemah = System Analysis and Management in Biomedical Systems, 2007, vol. 6, no. 2, pp. 395-400.

18. Korenevsky N. A., Korostelev A. N., Starodubtseva L. V., Serebrovsky V. V. Metod ocenki funkcional'nogo rezerva cheloveka-operatora na osnove kombinirovannyh pravil nechetkogo vyvoda [A method for assessing the functional reserve of a human operator based on the combined rules of fuzzy inference]. Biotekhnosfera = Biotechnosphere, 2012, no. 1 (19), pp. 44-49.

19. Korenevsky N. A., Bashir A. S., Gorbatenko S. A. Sintez gibridnyh nechetkih pravil dlya prognozirovaniya, ocenki i upravleniya sostoyaniem zdorov'ya v ekologicheski neblagopriyatnyh regionah [Synthesis of hybrid fuzzy rules for forecasting, assessment and management of health in ecologically unfavorable regions]. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta. Seriya: Upravlenie, vychislitel'naya tekhnika, informatika. Medicinskoe priborostroenie = Proceedings of the Southwest State University. Series: Control, Computer Engineering, Information Science. Medical Instruments Engineering, 2013, no. 4, pp. 69-73.

20. Al-Kasasbeh R. T., Alshamasin M. S., Korenevskiy N., Maksim I. Мethod of ergonomics assessment of technical systems and its influence on operators heath on basis of hybrid fuzzy models. Advances in Intelligent Systems and Computing, 2018, vol. 590, pp. 581-592.

21. Al-Kasasbeh R. T., Alshamasin M., Korenevskiy N., Kuzmin A., Ionescou F. Synthesis of fuzzy logic for prediction and medical diagnostics by energy characteristics of acupuncture points. JAMS Journal of Acupuncture and Meridian Studies, 2011, vol. 4, no. 3, pp. 175182.

22. Al-Kasasbeh R. T., Korenevskiy N. A., Alshamasin M. S., Korenevskya S. N., Al-Kasasbeh E. T., Maksim I. Fuzzy Model Evaluation of Vehicles Ergonomics and Its Influence on Occupational Diseases. Advances in Intelligent Systems and Computing, 2018, pp. 143-154.