08 GB Analysis of Variance-方差分析 - ANOVA.ppt

Copyright Route Six Sigma, LLC 2002 Analysis of Variance – ANOVA方差分析 - ANOVA What is Analysis of Variance Analysis of variance (ANOVA) is used to investigate and model the relationship between a dependent variable (Y) and one or more independent variables (X’s) that are qualitative (categorical). Analysis of variance extends the two-sample t-test for testing the equality of two population means to a more general null hypothesis of comparing the equality of more than two means, versus them not all being equal. An important tool for screening potential factors. 什么是方差分析 方差分析 (ANOVA)是用来探究和模型化依附变量(Y)与一个或更多独立变量(X’s) 之间关系的方法. 方差分析把 two-sample t实验比较两个总体均值数是否相等的方法拓展致利用一个一般的虚无假设,比较两个以上的均值, 假设它们并不都相等. 方差分析是一个揭示潜在因素的重要工具. The ANOVA Concept ANOVA is a tool for comparing multiple Means for Continuous Y. HO: m1=m2=m3=…=mL HA: Not all mi are equal ANOVA 概念 方差分析是一个用来为连续Y比较多个均值的工具. HO: m1=m2=m3=…=mL HA: 并不是所用 mi 都相等 Signal to Noise 信号和噪声 The Power of ANOVA Remember the limitation of “t” Test on the multiple machines. Open file slide.mtw ANOVA的功能(Power) 还记得 “t” Test实验对多种机器的限制条件么. 打开文件 file slide.mtw ANOVA Terms Response: The dependent variable (Y) Factor: An independent Variable (X) Level or Treatment: A discrete value or setting of a factor. Y=F(X) Balanced Data: There are equal numbers of samples or observations for each factor and treatment combination. The data above is balanced, since there are two observations for each level of X. Unbalanced Data: The number of observations is not the same for every factor and level combination. Y=F(X) The factor has two levels, 75 and 120, but there is only a single sample for 75 while there are three observations for 120. ANOVA 术语 回应 (response):依附变量(Y) 因素(factor): 独立变量 (X) 级别或处理(Level or Treatment): 离散数值或因素的构成 . Y=F(X) 平衡数据(Balanced Data):对每一个因素及处理的组合有相等数量的样本或观测. 上面的数据是平衡的,因为对每一X 的级别有两个观测结果. 非平衡数据(Unbalanced Data): 观测的数量对每一个因素和级别的组合不一致. Y=F(X) 因素有两个级别, 75 and 120, 对75只有一个样本 ，但对120有三次观测. Types of ANOVA(That we’ll use)