摘要
数理统计学教材中一般会介绍单因素方差分析和双因素方差分析,对多因素方差分析没有涉及。文章在双因素方差分析的基础上,将影响因素拓展为三个因子,并对三因素方差分析从模型的构建,方差平方和的分解和模型检验统计量的确定等进行了严格的数学推导,在此基础上,可以进一步构造多因素方差分析模型。
Univariate anova and bivariate anova are generally introduced in mathematical statistics textbooks, in which multivariate analysis of variance is not involved. Based on the two-factor variance analysis, this paper expands the influencing factors into three factors, and derives the three-factor variance analysis mathematically from the model construction, the decomposition of the sum of squares of variance, the determination of model test statistics, and so on, on the basis of which, the multi-factor variance analysis can be further constructed.
引文
[1]魏宗舒等.概率论与数理统计教程[M].北京:高等教育出版社,2004.
[2]茆诗松,王静龙,濮晓龙.高等数理统计[M].北京:高等教育出版社, 2006.
[3]茆诗松,周纪芗,陈颖.试验设计[M].北京:中国统计出版社,2004.
[4]刘文卿.实验设计[M].北京:清华大学出版社,2005.
[5]邢航.多因素方差分析中数学模型的建立与检验方法[J].电大理工,2008,235(2).