摘要
运用3种不同的方法所建立的多元回归预测模型,其回归预测值的准确度均达到了99 26%以上,其回归的预测值与实际值的方差分析检验均呈无差异性,P>0 50水平。说明3种不同的方法所建立的预测模型的预测值均具有显著性意义。在3种建模方法中,均有评价作用,且其评价目的也不尽相同。在体育运动实践中可依据研究的目的,选择建模方法,以提高体育预测研究的效益。
The paper, using statistical analysis of multiple regressions, suggests three different ways of modeling for anticipative results concerning athletic research, and obtains a valid anticipative value of 99.26%. And the value prescribed is closely significant compared with the actual One-way ANOVA (p>0.50). Thus the author thinks they are statistics-relevant in terms of value for result descriptions. The ways of modeling have their evaluative effects and only differ in their ends. The suggestion made might be helpful and effective in sports statistical analysis and modeling athletic research.
出处
《北京体育大学学报》
CSSCI
北大核心
2004年第12期1654-1656,共3页
Journal of Beijing Sport University
关键词
体育预测
多元回归
预测模型
比较研究
sports forecasting
multiple regression
anticipative modeling
ANOVA
weight factors