Process control and development of a decision support system for classifying information signals based on Markov models

The necessity of controlling the process of classifying information signals based on simple and two-connected Markov models is substantiated. The possibility of combining previously obtained models and a classification algorithm into a decision-making system in order to classify information signals (random processes) is shown according to the criterion of maximizing a posteriori probability. The article proposes a block diagram of the decision-making system, describes the developed software components that consistently implement both auxiliary and basic procedures that allow implementing previously synthesized Markov models and methods for evaluating their parameters, as well as a classification algorithm. The description of the possibility of learning the classification algorithm, both "with a teacher" and in the "self-learning" mode, is given, the volumes of samples of the observations provided by the studied signals for the formation of databases of Markov signal models, Markov models of signal classes are determined. The results of statistical simulation modeling of the dependence of the error probability on the size of the training sample are presented. Block diagrams of some software components of the decision support system are proposed. The results of the implementation of previously developed models, methods and algorithms in the form of software tools are considered, and the functionality of using these tools as part of a decision support system is shown. The results of calculations are presented, showing the adequacy of the solutions obtained and the functionality of the developed software tools. Conclusions are drawn about the possibility of using a decision support system in various subject areas, including when classifying the conditions of the patient's cardiovascular system according to the observed rhythmograms.

1. Kazakov V.A. Vvedenie v teoriyu markovskikh protsessov i nekotorye radiotekhnicheskie zadachi. Moscow: Sovetskoe radio; 1973. 232 p. (In Russ.).

2. Novikov D.A. Teorija upravlenija organizacionnymi sistemami. Moscow: Izdatel'stvo fiziko-matematicheskoj literatury; 2012. 604 p. (In Russ.).

3. Bocharov P.P., Pechinkin A.V. Teoriya veroyatnostei. Matematicheskaya statistika. Moscow: FIZMATLIT; 2005. 296 p. (In Russ.).

4. Glushkov A.N., Litvinenko Yu.V., Kalinin M.Yu. Markovskaya model' ritmogrammy. In: Vserossiiskaya konferentsiya «Intellektual'nye informatsionnye sistemy»: Intellektual'nye informatsionnye sistemy: trudy vserossiiskoi konferentsii, 23-24 June 2016, Voronezh, Russia. Voronezh: Voronezh State Technical University; 2016. P. 127–130. (In Russ.).

5. Kalinin M.Yu. Two-Dimensional Generalized Probability Model Radio Signals. Vestnik Voronezhskogo instituta vysokikh tekhnologii = Bulletin of the Voronezh Institute of High Technologies. 2021;15(1):16–18. (In Russ.). URL: https://vestnikvivt.ru/ru/journal/pdf?id=386 [Accessed 1st March 2024].

6. Glushkov D.A., Kalinin M.Yu., Litvinenko Yu.V. Analytical Two-Dimensional Probabilistic Radio Signal Model. Vestnik Voronezhskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of Voronezh State Technical University. 2023;19(1):57–61. (In Russ.).

7. Kalinin M.Yu., Choporov O.N. Learning the Algorithm for Classification of a Random Process. Vestnik Voronezhskogo instituta vysokikh tekhnologii = Bulletin of the Voronezh Institute of High Technologies. 2021;15(1):19–22. (In Russ.). URL: https://vestnikvivt.ru/ru/journal/pdf?id=387 [Accessed 1st March 2024].

8. Kalinin M.Yu., Choporov O.N., Bonch-Bruevich A.M. Classification of Random Signals Based on Their Doubly Connected Markov Models. Modelirovanie, optimizatsiya i informatsionnye tekhnologii = Modeling, Optimization and Information Technology. 2022;10(3). (In Russ.). URL: https://moitvivt.ru/ru/journal/pdf?id=1222. DOI: 10.26102/2310-6018/2022.38.3.017 [Accessed 1st March 2024].

9. Terelyanskii P.V. Sistemy podderzhki prinyatiya reshenii. Opyt proektirovaniya. Volgograd: VolgGTU; 2009. 329 p. (In Russ.).

10. Nikolenko S.I., Tulup'ev A.L. Samoobuchayushchiesya sistemy. Moscow: MTsNMO; 2009. 288 p. (In Russ.).

11. Kalinin M.Yu., Choporov O.N. Entropy Estimates of the Decision Statistics of the Classification Algorithm for Random Processes. Modelirovanie, optimizatsiya i informatsionnye tekhnologii = Modeling, Optimization and Information Technology. 2020;8(4). (In Russ.). URL: https://moitvivt.ru/ru/journal/pdf?id=881. DOI: 10.26102/2310-6018/2020.31.4.034 [Accessed 4th March 2024].