Predicting Models Of Olympic Games Based On Sequence And Multivariate Analysis

In this paper, the author analyzes the predicting models of the single sports scores and the Medals of Olympic Games, which based on the sequence analysis and factors analysis. And the author brings the statistical analysis methods of combination model, multi-dimensional misalignment model and nonlinear transformation model based on nuclear function into the prediction.First, the author studys the models of moving average, exponential smoothing and grey prediction which based on sequence analysis and the results show that the accuracy of each sports game to different single forecasting method. If we just choose one to forecast, the useful information provided by other models will be losted. So the author tryes to combine all the models by linear and nonlinear to forecast the sports scores of Olympic Games of2012.In addition, the medals of Olympic Games are concerned by all people. The author gets12economic factors which influence the number of medals to establish the linear multivariate regression model. Considering the limitation of the linear multivariate regression model, the author also establishes the nonlinear model to describe the12factors and the number of medals. Because nonlinear problems are solved by transform into linear problems. The author introduces the nonlinear model based on transfonnation by nuclear function and combines with partial least squares estimate to forecast the number of medals of2012Olympic Games.