摘要:本文旨在通过使用中超比赛中最主要的14支球队在2015赛季的统计数据,进行多元回归分析,构造胜率关于一些指标的回归方程,对于出现了的多重共线性问题,采取逐步回归的方法,剔除多余变量,最后得到了胜率关于一些重要指标的回归函数。然后根据以往赛季这些指标的实际值,再通过一次指数平滑法分析预测这些指标在2016赛季的预测值,并带入回归函数,得出各支球队在2016赛季的预测胜率,并由此进行排名,从而达到预测2016赛季中超比赛的结果。43689
毕业论文关键字:多元回归分析,多重共线性,逐步回归,一次指数平滑法,中超联赛
Abstract:In this paper, by using the statistical data in the 2015 competition season of the mainly 14 teams in Chinese Super League, multiple regression model are constructed to shows the linear relationship between winning probability and some indicators. But there is multicollinearity in this model, to solve this problem, the multiple stepwise regression method is used. Finally we get the regression function about winning probability and major indexes. Then, we get the major indexes values in last few competition seasons, use the single exponential smoothing method to predict the major indexes in 2016 competition season. Put it in the regression function, we can predict the winning probability about every team. So, the champion is obvious.
Key words: multiple regression analysis, multicollinearity, stepwise, single exponential smoothing, Chinese Super League
目录
1 绪论 1
1.1 选题背景和研究意义 1
1.3 本文结构 2
2 相关理论综述 4
2.1 多元回归分析 4
2.1.1 多元回归模型与回归方程 4
2.1.2 估计的多元回归方程 5
2.1.3 参数的最小二乘估计 5
2.2 回归方程的拟合优度 5
2.2.1 多重判定系数 5
2.2.2 显著性检验 6
2.3 多重共线性和逐步回归 7
2.4 一次指数平滑法 8
3 中超比赛的预测模型 9
3.1球队胜率的多元回归模型 9
3.1.1 结果分析 10
3.2 球队胜率逐步回归模型 11
3.3 指数平滑模型 12
4 结论与展望 16
5 致谢 17
6 参考文献