A SPATIOTEMPORAL MODEL PREDICTS THE DYNAMICS OF A BASKETBALL GAME
DOI:
https://doi.org/10.61841/ak6x8397Keywords:
Basketball game decision-making, Temporal point process, Markov decision process (MDP), LSTM-based architectureAbstract
Abstract—Our work as a spatiotemporal model predictive dynamic (STMPD) on a basketball game capable of high-level decision making . Our motivated and skilled experts solve complex decision-making problems at every time point during a game. The conditional intensity of a temporal point process can be learned based on a network, and the likelihood is maximised during training. A Fine-grained player locations started with the simplest case of modelling, where certain players tend to shoot the ball. They discretize the court into cells fine enough to capture all significant spatial variations in model for predicting dynamics in-game, e.g., shooting policy, passing, and who relies on Markov transition probabilities to propagate action through a possession. We propose a method for learning factored Markov decision problems . We use a method that captures the multi-modal behaviour of the players' truth ground and estimation trajectory Prediction is a method for predicting multimodal trajectory by learning a model that assigns probabilities through conducting temporal classification (resolution) (CTC) architectures by generalising away from actual time stamps to recreate the predicted trajectory. Our proposed method of contribution is as follows : i) The method is built on an LSTM-based architecture predicting multiple trajectories and our probabilities, and then we use training with a multimodal loss function when updating the best trajectories . ii) a discriminative learning method for automatically training models to predict near-term game events based on current game conditions . iii) learning factored Markov decision processes (FMDP) problems from domain exploration and expert assistance, ensuring convergence to near-optimal behaviour even after the agent has started to learn critical success factors. IV) An algorithm 1,2, detects changes in possession, pass, and shot players that have been incrementally learned for all components of FMDP, and algorithm 3, which guarantees convergence to near-optimal behaviour even when the agent begins unaware of factors critical to success. A group of players shooting probability to enhance a probabilistic matrix
References
Y. Yue, P. Lucey, P. Carr, A. Bialkowski and I. Matthews, “Learning Fine-Grained Spatial Models for Dynamic Sports Play Prediction, ” in procc: IEEE International Conference on Data Mining, 2014,pp.670-679.
B.Skinner and M. Goldman, Optimal strategy in basketball In Handbook of Statistical Methods and Analyses in Sports,2019 . [E-book]
M. Horton, J. Gudmundsson, S. Chawla and J. Estephan , “Automated classification of passing in football, ”in proc Intern Conference on Knowledge Discovery and Data Mining, Ho Chi Minh,Vietnam , 2015; pp.319–330.
D.Cervone, A.DAmour, L.Bornn,. and K.Goldsberry ,A multiresolution stochastic process model for predicting basketball possession outcomes, Journal of the American Statistical Association , vol .111, pp.585-599 , 2016.
A. Miller, L. Bornn, R. Adams, and K. Goldsberry, “Factorized point process intensities: A spatial analysis of professional basketball,” in International Conference on Machine Learning , vol 32, 2014.pp. 22-24.
V. Guralnik and J. Srivastava, “Event detection from time series data,” in Proc: fifth ACM SIGKDD international conference on Knowledge discovery and data mining. New York, NY, USA:ACM Press, 1999, pp. 33–42.
M. Salmenkivi and H. Mannila, “Using Markov Chain Monte Carlo and dynamic programming for event sequence data,” Knowledge and Information Systems, vol. 7, no. 3, pp. 267–288, 2005.
S. L. Scott and P. Smyth, “The Markov modulated Poisson process and Markov Poisson cascade with applications to web traffic data,” Bayesian Statistics, vol. 7, 2003, pp. 671–680.
A.Alahi, K.Goel, V.Ramanathan, A.Robicquet, L.Fei and S.Savarese, “Social lstm: Human trajectory prediction in crowded spaces, ” in proc: IEEE conference on computer vision and pattern recognition, 2016,pp. 961–971.
S.Haddad and S.Lam, “ Self-growing spatial graph networks for pedestrian trajectory prediction" in Proc. The IEEE Winter Conference on Applications of Computer Vision, 2020, pp. 1151–1159.
A. Graves, S. Fern´andez, F. Gomez, and J. Schmidhuber, “Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks,”in ICML, Garching, Munich:Germany ,2006.
L.Lamas, J.Barrera, G.Otranto, and C.Ugrinowitsch, invasion team sports: strategy and match modeling
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Journal of Advance Research in Applied Science (ISSN 2208-2352)
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Terms & Condition
Submission -
Author can submit the manuscript through our online submission process or email us at the designated email id in contact details.
The other mode of submission not accepted than online and email.
Before submission please read the submission guidelines.
NN Publication accepts only article submitted in pdf/doc/docx/rtf file format. Another format except given file formats will no be considered .
Author will be responsible for the error mistakes in the submission files. The minor changes can be done without any cost after publication. But for major changes NN Publication may charges you the editing charges.
Publication (Online) -
The online publication is scheduled on last date of every month, but it can be delayed by 24 to 48 hours due to editorial process if huge number of articles comes to publish in single issue.
Automatic notificatation email will be sent to the all users on publication of an issue, so its author’s duty to check their email inbox or SPAM folder to get this notification.
After publication of article author can not withdraw their article.
If editor’s found any issue after publication of article then the NN Publication have the authority to remove the article from online website.
No refund will be provided after online publication of article.
Publication (Print) -
The print copy publication are sent as per the author’s request after 2 weeks of online publication of that issue.
NN Publication will ship the article by India Post and provide the consignment number on dispatch of print copy.
NN Publication follows all the guidelines of delivery provided by IndiaPost and hence not responsible for delay in delivery due to any kind of reasons.
Refund of hard copy will not be provided after dispatch or print of the journal.
NN Publication will be responsible for raise a complain if there is any issue occurs in delivery, but still will not be responsible for providing the refund.
NN Publication will be responsible to resend the print copy only and only if the print copy is lost or print copy is damaged in delivery / or there is delay more than 6 months.
According to India Post the delivery should be completed with in 1-3 weeks after dispatch of articles.
Privacy Policy-
NN Publicationl uses the email ids of authors and editors and readers for sending editorial or publication notification only, we do not reveal or sell the email ids to any other website or company.
