S.I. Loginov, professor, Dr.Biol.
A.A. Egorov, Ph.D.
V.A. Ermakov, Honored coach of the RF in judo
Surgut State University Khanty-Mansi Autonomous Region Yugra, Surgut

Key words: elite sport, computer technology, informational support, data mining.

Introduction. Nowadays information (digital) technology is an indispensable attribute of sports, health and fitness and biomedical practice [1]. Electronic document management is being carried out and statistical reports and databases are being prepared with the help of specially designed software in most of sports, educational and medical institutions [8]. A more complex product of digital technology (DT) is information support of sports events and games in professional team sports leagues [6]. Data mining systems (MedMining) are of particular attention in the spheres of elite sport and biomedicine. However, notwithstanding the general remarkable success of DT, the application of data mining techniques in the field of sport is still limited [1, 3]. They mainly concern the issues such as evaluation of game strategy, forecast of efficiency of training load and risk of injury, analysis of team and individual actions in various, mostly team, sports [5, 7]. The number of studies devoted to the development of data mining systems in sport in general, and combat sports in particular, is very limited [1, 9]. The researchers of the Laboratory of Biomechanics and Kinesiology, the Department of Automated Information Processing and Management Systems, LLC "Design office "Automated systems and system analysis" under Surgut State University, are implementing a joint project involving the development of the program for collection, processing, analysis and management of biomedical indicators of athletes of the Children's Sports School of the Olympic Reserve in Surgut. In the present paper we examine the main prerequisites for the creation of the data mining system and possible aspects of its implementation during the long-term training of the Olympic reserve and elite judokas.

The purpose of the study is to create an information system for automated processing, storage, analysis, monitoring and management of training and competitive activities of sports schools of the Olympic reserve, judo module, based on the original mathematical and statistical models.   

Results and discussion. The process of training is a classic control system. A trainer and an athlete act as subjects (actors) of the system. The athlete is an object of control, and the trainer is a subject that makes decisions, i.e. the one providing control action. At the same time the entity “athlete” has both an internal state and a state determined by the environment. The environment is a plurality of determinants specific for a particular society. It is probabilistic in nature and reflects the ambiguity of nature. The output of the system is a set of indicators and individual achievements of an athlete. In general terms the diagram is shown in Figure 1.

This diagram reflects the essence of individual training sessions. In practice, the training process is most often of a group nature. The object of control then is presented by masses of athletes (5-20 athletes) wherein each athlete has an impact y`₀(t) on the remaining part of the masses of athletes (Figure 2).

The presence of seven or more objects of control in a system indicates its high complexity. At the same time each object of control, in its turn, is also a complex system that is difficult to be formalized. The number of athlete’s output parameters at the moment of time T is represented by a hundred or more numerical values. Thus, the managing object (a trainer) must handle thousands of indicators when making a decision. 

Subject area model. A subject area forms a plurality of entities and relationships between them. All functions of the information system are subject to the logic shown in Figure 3. The structure reflects the main entities of the subject area of the training process of the Children's Sports School of the Olympic Reserve in the sports discipline of judo. The relationships between these entities are shown as interrelations of objects.

The subject area model describes processes taking place within the framework of the training process (TP), organization and management of sports competitions, shows the relations between trainers – groups of athletes – fights – athletes.

An athlete is the key entity of the model. In fact, this entity is an independent model describing a lot of group and integral parameters of a real athlete. These parameters include: track records of an athlete (achievements, Figure 1); map of tactical and technical infrastructure; attendance report, basic data about the athlete; contact details of the athlete and his parents; medicobiological passport data of the athlete; a list of disease risks; indicators of overall (PhC) and special physical conditions (SPhC); morphological and functional assessment indicators; sociological and psychological indicators.   

Additional processes implemented via the information system include the following: management of a sports school; management of the system users; management of trainers; sports groups management; management of the training process planning; reporting subsystem management; survey subsystem management; management of morphological and functional assessment techniques; management of the process of transfer tests; management of the school’s competitive activity planning.

Information system structure. The information system is built based on the three-tier architecture and contains the following program components: an application server – implements the major business logic of interaction of the objects in the system, ensures interaction of the server and client applications based on TCP/IP protocol, transmits queries to the database server and processes its response. Database server (DBS) – operates with the database management system. All system data are stored in the database. Client – an application running in the client hardware browser (Figure 4).  

The information system consists of the following main modules. The system interface is implemented on MVC 5 platform and is a thin client made in the form of a Web application using HTML5. The module “software” implements the functionality of data mining. This module contains a set of statistical data analysis tools, a neural network device, a device for solving differential equations.  

The database access module is implemented on platforms Entity Framework 6.0 and Framework 4.5. It contains an object-oriented domain model, provides access to the database management system and controls the dataflow between the modules and the database. The components module contains sets of advanced components that provide user interface control. The charts and graphs module ensures visualization of intermediate data and data analysis results. The reports module generates reports about results of activities of sports schools and analytical reports on the results of the training process. 

Conclusion. The information system presented herein is designed to organize and manage the training process and competitive activity in a judo school based on computer technology as well as data processing automation. The system will help classify athletes and predict their further improvement and effectiveness by means of fuzzy sets and big data mining.   

References

1. Epishev, V.V. Sistema intellektual’nogo analiza dannykh fiziologicheskikh issledovaniy v sporte vysshikh dostizheniy (Intelligent data analysis of physiological research in elite sport) / V.V. Epishev, A.P. Isaev, R.M. Miniakhmetov, A.V. Movchan, A.S. Smirnov [et al.] // Vestnik YuUrGU. Seriya “Vychislitel’naya matematika i informatika” (Series «Computational Mathematics and Computer Science»). – 2013. – Part 2, № 1. – P. 44–54.
2. Oplachko, E.S. Oblachnye tekhnologii i ikh primenenie v zadachakh matematicheskoy biologii (Cloud technologies and their application in mathematical biology) / E.S. Oplachko, D.M. Ustinin, M.N. Ustinin // Matematicheskaya biologiya i bioinformatika. – 2013. – Part 8, № 2. – P. 449–466.
3. Highlight, M. A Review of Data Mining Techniques for Result Prediction in Sports / M. Highlight, H. Rastegari and N. Nourafza // ACSIJ Advances in Computer Science: an International Journal, 2013. V.2, Issue 5, N 6. – November 2013 ISSN: 2322-5157 www.ACSIJ.org.
4. Lavrac, N. Intelligent Data Analysis in Medicine. / N. Lavrac, E. Keravnou, B. Zupan. In: Kent A, editor. In Encyclopedia of Computer Science and Technology. (New York) 2000. – V. 42. – P. 113–157.
5. McCullagh, J. Data Mining in Sport: A Neural Network Approach / J. McCullagh // International Journal of Sports Science and Engineering. 2010. – V.04. – N. 03. – P. 131–138.
6. Miljkovic, D. The use of data mining for basketball matches outcomes prediction / D. Miljkovic, L. Gajic, A. Kovacevic, Z. Konjovic. IEEE 8^th International Symposium on intelligent and informatics, Subotica, Serbia. 2010. – P. 309–312.
7. Schumaker, R.P. Sports knowledge management and data mining. / R.P. Schumaker, O.K. Solieman, H. Chen. // Annual Review of Information Science and Technology. 2010. – V. 44. – N 1. – P. 115–157.
8. Shivade, C. A review of approaches to identifying patient phenotype cohorts using electronic health records / C. Shivade, P. Raghavan, E. Fosler-Lussier, P.J. Embi, N. Elhadad, S.B. Johnson, A.M. Lai // J. Am. Med. Inform. Assoc. 2014. – V. 21, N 2. – P. 221–230. doi: 10.1136/amiajnl-2013-001935.
9. Zhu, L. Introduction to medical data mining. / L. Zhu, B. Wu, C. Cao // Sheng Wu Yi Xue Gong Cheng Xue Za Zhi. – 2003. –V. 20, N 3. – P. 559–562.

Corresponding author: apokin_vv@mail.ru