Predictive Modeling of Olympic Medal Using Adaptive Kernel Density Estimation

Considering that the traditional Olympic medal predicting method cannot deal with non-parametric statistics in medal predicting, this paper uses a new Adaptive KDE mode to predict and analyze the gold medal list of 2028 Olympics. The predicted value distribution is not parametric statistic and more flexible than ordinary regression models. Using Python simulation establishes an arithmetic calculation mode, generating enough numbers of samples to increase the reliability of prediction. Through simulation experiment, it is verified that the winning or losing probability predicting accuracy is high; at the same time, when other data sets are used for testing, the effect obtained from fitting through this algorithm model will not decrease but only gradually tend toward the best condition; its fit effect has certain research significance for Olympic competition prediction in sports training.