Evaluating Momentum-Weighted LSTM Models for Predicting Tennis Match Outcomes
DOI:
https://doi.org/10.54097/zhzwq146Keywords:
Long Short-Term Memory, Cross-Validation, Momentum, Tennis match prediction, Real-time performance.Abstract
In the realm of sports competition, particularly in skill- and strategy-centric sports like tennis, momentum stands out as a pivotal factor influencing match outcomes. Conventional match analysis methods often fall short in capturing the nuanced shifts in momentum and their profound impact on players' psychological states and performances. To precisely forecast momentum changes and their repercussions, we introduce two integrated concepts: momentum and a player's point-winning probability. Weight calculations are optimized through cross-validation, and a model predicting match trends is formulated using the LSTM neural network theory. The results underscore a direct correlation between momentum swings and players' scores, significantly shaping match outcomes. Evaluation metrics reveal the model's high accuracy (87.06%) and minimal loss (0.6655) on the test set, attesting to its robust predictive capabilities. This research not only advances our understanding of momentum dynamics in sports but also showcases a methodological approach with potential applications across various disciplines, fostering a more nuanced comprehension of competitive dynamics.
Downloads
References
Fayomi, A., et al., Forecasting Tennis Match Results Using the Bradley-Terry Model. International Journal of Photoenergy, 2022. 2022.
Wu, E. and H. Koike. Futurepong: Real-time table tennis trajectory forecasting using pose prediction network. in Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems. 2020.
Hochreiter, S. and J. Schmid Huber, long short-term memory. Neural computation, 1997. 9 (8): p. 1735 - 1780.
Zhang, Q., et al., Sports match prediction model for training and exercise using attention-based LSTM network. Digital Communications and Networks, 2022. 8 (4): p. 508 - 515.
Zhao, Y., et al., Applying deep bidirectional LSTM and mixture density network for basketball trajectory prediction. Optik, 2018. 158: p. 266 - 272.
Zhang, L., et al., Improved Dota2 lineup recommendation model based on a bidirectional LSTM. Tsinghua Science and Technology, 2020. 25 (6): p. 712 - 720.
Bates, S., T. Hastie and R. Tibshirani, Cross-validation: what does it estimate and how well does it do it? Journal of the American Statistical Association, 2023: p. 1 - 12.
Staudemeyer, R.C. and E.R. Morris, Understanding LSTM--a tutorial into long short-term memory recurrent neural networks. arXiv preprint arXiv:1909.09586, 2019.
Yu, Y., et al., A review of recurrent neural networks: LSTM cells and network architectures. Neural computation, 2019. 31 (7): p. 1235 - 1270.
Li, S., et al. Independently recurrent neural network (indrnn): Building a longer and deeper rnn. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2018.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Highlights in Science, Engineering and Technology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
