The Prediction and Analysis of Medal Counts Based On ALRWI-BPLP

Abstract

Accurately predicting Olympic medal counts is crucial for evaluating national sporting strength and optimizing resource allocation. Existing prediction methods often fail to fully leverage rich historical information and perform poorly under data-sparse conditions. To address this, this paper proposes a novel hybrid prediction model: For events with sufficient data, a Backpropagation (BP) neural network employing adaptive learning rates (ALR) and Xavier weight initialization (WI) is utilized, incorporating features such as participant numbers from the previous two Games and host nation status for refined prediction. For events with insufficient or no historical data, a combined model primarily based on linear programming (LP), supplemented with Gaussian white noise correction, is adopted. This strategy effectively addresses challenges including zero-value interference in historical data, and data volume disparities, while enhancing prediction robustness. Applied to Olympic data from 1896 to 2024, the model successfully predicted the total medal counts for nations at the 2028 Los Angeles Olympics. The results project: the United States (155 medals), China (82 medals), and France (76 medals) will secure the top three positions on the medal table. Compared to 2024, the medal counts of the United States, France, Australia, and others are projected to increase, while those of China, Great Britain, the Netherlands, and others are expected to decline. Concurrently, forecasting results separately for each sport within a country provides clear and intuitive insights into the nation's degree of reliance on specific sporting events. This enables National Olympic Committees (NOCs) to form expectations regarding potential medal distributions in future Games, thereby facilitating more rational resource allocation to maximize performance outcomes. The findings of this study thus offer a scientific basis and practical reference for Olympic project analysis and strategic decision-making by NOCs regarding resource allocation optimization.