Prediction of Olympic medals based on GA-BP and logistic regression model

Predicting the number and distribution of medals in the future Olympic Games has become a noticeable topic. However, it is not easy to predict the number of Olympic medals. We need to comprehensively consider historical data, athletes' performance, host country effect and other factors from multiple dimensions. This paper adopts GA-BP algorithm model, which is an optimization algorithm combining genetic algorithm (GA) and back propagation neural network (BPNN). It uses the global search ability of genetic algorithm to optimize the weight and bias parameters of BP neural network, so as to improve the training efficiency and prediction performance of neural network. By estimating the number of gold medals and total medals of each country in the Olympic Games, and verifying the accuracy of the model, the medal list of the 2028 Los Angeles Olympic Games is finally predicted. Based on the synthetic control model, this paper selects two countries, Estonia and China, as the research object. By constructing a virtual control group and two experimental groups, the change of the number of medals in different periods and the change of the number of medals with or without coaches in the same period were analyzed, and the difference between the actual medal growth and the synthetic control group (no coach effect team) was compared. The results show that the number of medals won by Estonia and China with a head coach is higher than that without a head coach. Under the guidance of excellent coaches, Estonia won 1 gold medal and 2 bronze medals in 1992, indicating that the great head coach effect has a significant role in promoting the performance of athletes. This paper explores the valuable opinions of the Olympic committee decision-making, reveals the key factors behind the distribution of medals, optimizes the allocation of national strategic resources, and predicts the performance of countries in the future Olympic Games. This research will help to improve the accuracy of the model more comprehensively, better understand the medal trend of the Olympic Games, and provide a scientific basis for the NOC to achieve better results in the future Olympic Games.

预测未来奥运会奖牌的数量与分布已成为备受关注的议题。然而，奥运会奖牌数量的预测并非易事，需从多维度综合考量历史数据、运动员竞技表现、东道主效应等诸多因素。
本文采用GA-BP算法模型，即结合遗传算法（Genetic Algorithm，GA）与反向传播神经网络（Back Propagation Neural Network，BPNN）的优化算法，借助遗传算法的全局搜索能力优化BP神经网络的权重与偏置参数，以此提升神经网络的训练效率与预测性能。通过对奥运会各国金牌数与总奖牌数进行预估，并验证模型的预测精度，最终完成2028年洛杉矶奥运会奖牌榜的预测。
本文基于合成控制模型（Synthetic Control Model），选取爱沙尼亚与中国作为研究对象，通过构建虚拟对照组与两个实验组，分析不同时期的奖牌数量变化，以及同期配备主教练与未配备主教练时的奖牌数量差异，并对比实际奖牌增长情况与合成对照组（无主教练效应队伍）的差值。
研究结果表明，配备主教练的爱沙尼亚与中国所获奖牌数均高于未配备主教练的情况。在优秀主教练的指导下，爱沙尼亚于1992年斩获1枚金牌与2枚铜牌，足见优秀主教练效应对运动员竞技表现具有显著的促进作用。
本文旨在为奥委会决策提供有价值的参考意见，揭示奖牌分布背后的关键影响因素，优化国家战略资源配置，并预测各国在未来奥运会中的竞技表现。本研究有助于更全面地提升模型预测精度，更好地把握奥运会奖牌分布趋势，为国家奥委会（National Olympic Committee，NOC）在未来奥运会中取得更优竞技成绩提供科学依据。