A study of tennis match trend prediction based on momentum assessment and heuristic HMMs

Xuanjun He; Changxu He; Fuxian Ye

doi:10.54097/jj1vs437

Authors

Xuanjun He
Changxu He
Fuxian Ye

DOI:

https://doi.org/10.54097/jj1vs437

Keywords:

Sports Momentum, Heuristic HMM, Match Prediction.

Abstract

In competitive tournaments, there exists a currently unexplored factor, momentum, whose role affects the trend of the game to a certain extent and even produces surprising fluctuations, which is particularly evident in tennis.This paper proposes a feasible research programme for the problem of momentum change and trend prediction in tennis matches. Firstly, hundreds of matches of Wimbledon Tennis Championships were used as the data source, and the assessment indexes related to momentum were filtered and mathematical models were established through analysis and processing, and then the coefficient settings were optimised by using the self-looping simulated annealing algorithm. The final model better describes changes in match conditions, with a model fit of over 85% for 87% of the matches.At the same time, in order to verify the existence of correlation between the momentum and the direction of the game, this paper randomly selects a number of games, the first half of the event as a dataset for the momentum model training, and uses the heuristic HMM to predict the subsequent changes and compare them with the scoring difference in the second half of the actual game, and the test result is that the two have basically the same trend, and therefore it can be proved that the momentum change curve can be used as one of the indicators of the prediction of the game trend.

Downloads

Download data is not yet available.

References

C.Y. Huang. Research on tennis match performance prediction model based on BP neural network [D]. Fujian Normal University, 2022.

JIANG Hai-Fu, YU Hua-Long, WEI Lei. Research on the analysis of influencing factors of soccer game based on fuzzy gray correlation analysis [J]. Computer and Digital Engineering, 2023, 51(03): 555-560.

Qi Hang. Discriminant analysis and outcome prediction of winning factors in ATP men's singles matches [D]. China University of Geosciences (Beijing), 2014.

YANG Longfei, LIU Qingyao, ZHANG Jiangnan. Research and analysis on analyzing and predicting the results of men's singles badminton matches in China and the structure of winning and losing[C]//China Wisdom Engineering Research Association, China Bandi Association, Guangdong Physical Fitness Association. Proceedings of the 10th China Physical Training Science Conference (in Chinese). [Publisher unknown], 2023: 5.

Rong Zhang. Result prediction and player analysis of tennis matches [D]. Yunnan University, 2023.

WU Jun-Ying, LU Xin, LIU Hong et al. Ultra-short-term multiregional power load forecasting based on Spearman-GCN-GRU model [J/OL]. China Electric Power, 1-9 [2024-03-16].

Kang Wenxuan. Urban delivery path planning for shared bicycle based on simulated annealing algorithm [J]. Technology and Innovation, 2022(13): 104-106+109.

HAN J X, MA M Y, WANG K. Product modeling design based on genetic algorithm and BP neuralnetwork []. Neural Computing and Applications, 2021, 33(9): 4111-4117

Ni Xuan, Chen Yichen. ARIMA model combined with LSTM neural network to establish a prediction model for drug purchasing decision [J]. Heilongjiang Science, 2024, 15(02): 76-78.

Wang Xinwu. Research on online video personalized recommendation based on Hidden Markov Model [J]. Digital Communication World, 2023(09): 1-4.

REN Yongfeng, XUE Yu, YUN Pingping et al. Markov prediction of multi-objective optimized energy storage system for smoothing power fluctuation in wind farms [J]. Power System Automation, 2020, 44(06): 67-74.

Zhao Yue, Zhang Xueyong. Research on the application of Bayesian regularization in spherical equivalent source method [J]. Journal of Anhui University of Engineering, 2023, 38(02): 89-94.