Decoding Olympic Glory: A Data-Driven Approach to Medal Predictions and Strategic Insights

Peijun Dong, Mingtao He, Zengrui Xu
Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing, PMLR 278:125-131, 2025.

Abstract

The objective of this study is to investigate the potential of data science and machine learning to find optimal performance levels for athletes and maximize strategies for national improvement. Utilizing Olympic data from 1986 to 2024, it implements the Fusion Medal Prediction Model (FMPM) to assist decision-making for athletes and coaches. Initially, we establish the XG-Prophet Model to forecast the 2028 Olympic medal table with MAE equals to 1.09/0.95/1.12/2.24 for gold/silver/bronze/total medals respectively. Additionally, GRU-ARIMA + XGBoost (Fusion Learning, ROC-AUC: 0.917) to identify the first winner. Furthermore, we explore medal distributions based on event types, employing K-means clustering to observe different contributions by country and finds that field and swimming events are key across all countries but with varying importance. An examination of event selection on a country-to-country basis through correlation shows that if the host country selects stronger events, even potentially favoring those in which they can excel, their numbers of medals naturally increase.

Cite this Paper

BibTeX
@InProceedings{pmlr-v278-dong25a, title = {Decoding Olympic Glory: A Data-Driven Approach to Medal Predictions and Strategic Insights}, author = {Dong, Peijun and He, Mingtao and Xu, Zengrui}, booktitle = {Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing}, pages = {125--131}, year = {2025}, editor = {Zeng, Nianyin and Pachori, Ram Bilas and Wang, Dongshu}, volume = {278}, series = {Proceedings of Machine Learning Research}, month = {25--27 Apr}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v278/main/assets/dong25a/dong25a.pdf}, url = {https://proceedings.mlr.press/v278/dong25a.html}, abstract = {The objective of this study is to investigate the potential of data science and machine learning to find optimal performance levels for athletes and maximize strategies for national improvement. Utilizing Olympic data from 1986 to 2024, it implements the Fusion Medal Prediction Model (FMPM) to assist decision-making for athletes and coaches. Initially, we establish the XG-Prophet Model to forecast the 2028 Olympic medal table with MAE equals to 1.09/0.95/1.12/2.24 for gold/silver/bronze/total medals respectively. Additionally, GRU-ARIMA + XGBoost (Fusion Learning, ROC-AUC: 0.917) to identify the first winner. Furthermore, we explore medal distributions based on event types, employing K-means clustering to observe different contributions by country and finds that field and swimming events are key across all countries but with varying importance. An examination of event selection on a country-to-country basis through correlation shows that if the host country selects stronger events, even potentially favoring those in which they can excel, their numbers of medals naturally increase.} }
Endnote
%0 Conference Paper %T Decoding Olympic Glory: A Data-Driven Approach to Medal Predictions and Strategic Insights %A Peijun Dong %A Mingtao He %A Zengrui Xu %B Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing %C Proceedings of Machine Learning Research %D 2025 %E Nianyin Zeng %E Ram Bilas Pachori %E Dongshu Wang %F pmlr-v278-dong25a %I PMLR %P 125--131 %U https://proceedings.mlr.press/v278/dong25a.html %V 278 %X The objective of this study is to investigate the potential of data science and machine learning to find optimal performance levels for athletes and maximize strategies for national improvement. Utilizing Olympic data from 1986 to 2024, it implements the Fusion Medal Prediction Model (FMPM) to assist decision-making for athletes and coaches. Initially, we establish the XG-Prophet Model to forecast the 2028 Olympic medal table with MAE equals to 1.09/0.95/1.12/2.24 for gold/silver/bronze/total medals respectively. Additionally, GRU-ARIMA + XGBoost (Fusion Learning, ROC-AUC: 0.917) to identify the first winner. Furthermore, we explore medal distributions based on event types, employing K-means clustering to observe different contributions by country and finds that field and swimming events are key across all countries but with varying importance. An examination of event selection on a country-to-country basis through correlation shows that if the host country selects stronger events, even potentially favoring those in which they can excel, their numbers of medals naturally increase.
APA
Dong, P., He, M. & Xu, Z.. (2025). Decoding Olympic Glory: A Data-Driven Approach to Medal Predictions and Strategic Insights. Proceedings of 2025 2nd International Conference on Machine Learning and Intelligent Computing, in Proceedings of Machine Learning Research 278:125-131 Available from https://proceedings.mlr.press/v278/dong25a.html.

Related Material