Нейромережеве моделювання та прогнозування актуалізації кіберспортивної індустрії на світовому рівні

Abstract

В статті досліджені питання зростання зацікавленості користувачів інтернет до галузі кіберспорту. Побудовано математичні моделі нейронної мережі актуалізації кіберспортивної індустрії на світовому рівні. Виконано прогнозування на основі радіальних базисних функцій та багатошарового персептрону.

E-sports today is a relatively young and growing industry, which is increasingly attracting the attention of the global community of connoisseurs of competitive video games, as well as attracting ordinary fans of sporting events. This explains the active sponsorship of world-famous brands of tournaments in various disciplines, as well as the importance of this sport, which can now be compared with traditional ones. The article considers the branching of electronic games from the gaming industry into the category of sports, which has considerable social importance. Simultaneously with the development of the e-sports industry, the number of scientific studies on this topic is growing, there are publications on economic issues, legal, psychological, as well as mathematical direction. The article examines the growing interest of Internet users in the field of e-sports. Based on the statistics of search queries in the Internet browsers of the Google Trend service for 17 years, mathematical models of the neural network of actualization of the e-sports industry at the global level have been built. The proposed approach to the creation of neural networks is based on the use of network models based on radial basis functions RBF and multilayer perceptron MLP. The creation of a neural network using radial basis functions is based on the use of the RBFT algorithm. The construction of a neural network using a multilayer perceptron is based on the use of the BFGS algorithm. Created models of neural networks are described by a set of system elements such as: network architecture (number of layers and hidden neurons), performance and errors (training, test), network learning algorithm, error functions, active hidden and active source neurons. The forecasting of the actualization of the e-sports industry at the world level was performed, which confirmed the hypothesis of gradual growth of the studied process.