Game method of ontology clustering

2019;
: pp. 26-39
1
Lviv Polytechnic National University, Information Systems and Networks Department
2
Lviv Polytechnic National University, Information Systems and Networks Department
3
Lviv Polytechnic National University

In this paper the important problem of ontology clustering is considered with the purpose of optimization of intelligent data processing in conditions of uncertainty caused by inaccuracy or incompleteness of data in the subject area. The clustering of ontologies is the process of automatic splitting of a set of ontologies into groups (clusters) based on their similarity degree. For the resolution of this problem it is necessary to adopt the set of measures for the affinity of ontologies, to choose or develop an algorithm of clusterization and to execute the thorough interpretation of clusterization results.
For the clustering of ontologies in conditions of uncertainty, it is proposed to use a stochastic game method. A repetitive stochastic game consists in the implementation of a controlled random process for selecting clusters of ontologies. To this effect, the intelligent agents, assigned to ontology, randomly, simultaneously and independently choose one of the clusters at discrete moments of time. For agents that have selected a cluster, the current measure of similarity of ontologies is calculated, which takes into account the proximity of concepts, attributes, and relationships between concepts. This measure is used to adapt the recalculation of mixed player strategies. Thus, the probability of selection is increased for clusters having the composition, which led to the growth of the ontologies similarity degree.
During the repetitive game, agents will form vectors of mixed strategies that will maximize the averaged measures of similarity to clusters of ontologies.
To solve the problem of game clusterization for ontologies, an adaptive Markovian recurrent method was developed based on stochastic approximation of a modified complementary slackness condition, valid at the points of the Nash equilibrium. The proposed game method has filtering properties for spikes in the input data and practically does not depend on the law of distribution of random noises.
The computer modeling confirmed the possibility of using a stochastic game model for clustering ontologies, taking into account uncertainty factors. Convergence of the game method is ensured by observing the fundamental conditions and restrictions of stochastic optimization. The reliability of experimental studies is confirmed by the repeatability of results obtained for various sequences of random variables.
The results of the work could be used to solve the problems of intellectual data analysis, to eliminate duplication of information in knowledge bases, to reduce uncertainty within the cluster of ontologies, to identify the novelty of information, to organize high-level semanticinteraction between agents in the course  of executing their common task.

1. Wooldridge, M. (2009). An Introduction to Multiagent Systems. United Kingdom: John Wiley & Sons.

2. Rogushina, Yu. V. (2018). Theoretical principles of application of ontologies for systematization of WEB resources. Problems of programming, 2-3, 197-203.

3. Hashemi, P., Khadivar, A., Shamizanjani, M. (2018). Developing a domain ontology for knowledge management technologies. Online Information Review, 42 (1), 28-44. https://doi.org/10.1108/OIR-07-2016-0177

4. Dovgy, S. O., Velychko, V. Yu., Globa, L. S., at al. (2013). Computer ontologies and their use in the educational process. Theory and practice: Monograph. Kyiv: Institute of Gifted Child.

5. Burov, E. V., Pasichnyk, V. V. (2015). Software systems based on ontological task models. Bulletin of the Lviv Polytechnic National University. Series: "Information Systems and Networks", 829, 36-57.

6. Berko, A., Alieksieiev, V. (2018) A method to solve uncertainty problem for Big Data sources. Proceedings of the 2018 IEEE 2nd International Conference on Data Stream Mining and Processing (DSMP). Lviv, Ukraine, August 21-25, 32-37. https://doi.org/10.1109/DSMP.2018.8478460

7. Mirkin, B. G. (2005). Clustering for Data Mining. A Data Recovery Approach. CRC Press. https://doi.org/10.1201/9781420034912

8. Batet, M. (2011). Ontology-based semantic clustering. AI Communications, 24 (3), 291-292.

9. Zaychenko, Yu. P., Gonchar, M. A. (2007) Fuzzy methods of cluster analysis in problems of automatic classification in economics. Bulletin of the NTU of Ukraine "Kyiv Polytechnic Institute". Informatics. Series: "Management and Computing", 47, 198-206.

10. Bodiansky, E. V., Kolchigin, B. V., Volkova, V. V., Pliss, I. P. (2013). Adaptive fuzzy clustering of data based on the Gustafson-Kessel method. Control systems and machines, 2, 40-46.

11. Lytvyn, V. V., Vysotska, V. A., Dosyn, D. G., Girnyak, M. G. (2015). Development of methods and means of constructing intelligent systems for processing information resources using the ontological approach. Bulletin of the Lviv Polytechnic National University. Series: "Information systems and networks", 832, 295-314.

12. Aleman, Y., Somodevilla, M. J. (2017). A proposal for domain ontological learning. Research in Computing Science, 133, 63-70.

13. Lytvyn, V. V. (2011). Intelligent search agents of relevant precedents based on adaptive ontologies. Mathematical Machines and Systems, 3, 66-72.

14. Ovdii, O. M., Proskurina, G. Yu. (2004). Ontology in the context of information integration: concepts, methods and construction tools. Problems of Programming, 2-3, 353-365.

15. Chistyakova, I. S. (2014). Ontology engineering. Software Engineering, 4 (20), 53-68.

16. Slimani, T. (2015). Ontology Development: A Comparing Study on Tools, Languages and Formalisms. Indian Journal of Science and Technology, 8 (24), 1 - 12. https://doi.org/10.17485/ijst/2015/v8i1/54249

17. Krjukov, K. V., Pankova, L. A., Pronina, V. A., Sukhoverov, V. S., Shipilina, L. B. (2010). Measures of semantic proximity in ontology. Problems of Management, 5, 2-14.

18. Neyman, A., Sorin, S. (2012). Stochastic Games and Applications. Springer Science & Business Media.

19. Nazin, A. V., Poznyak, A. S. (1986). Adaptive Choice of Variants: Recurrence Algorithms. Moscow: Science.

20. Petrosjan, L.A., Mazalov, V. V. (2007). Game Theory and Application. New York: Nova Science Publishers.

21. Neogy, S. K., Bapat, Ravindra B., Dubey, Dipti. (2018). Mathematical Programming and Game Theory. Springer. https://doi.org/10.1007/978-981-13-3059-9

22. Kushner, H., Yin, G. G. (2013). Stochastic Approximation and Recursive Algorithms and Applications. Springer Science & Business Media.