问题描述
我需要一些帮助来执行R中的N向ANOVA，以捕获不同因素之间的相互依存关系.在我的数据中，大约有100个不同的因素，我正在使用以下代码执行ANOVA.
I need some help in performing N-way ANOVA in R to capture inter dependencies among different factors. In my data, there are around 100 different factors and I am using the following code to perform ANOVA.
model.lm<-lm(y~., data=data)
anova(model.lm)
据我所知(可能是我错了)，它仅对每个因子执行1向ANOVA.由于某些原因，我需要在所有100个组之间(即从x1到x100)执行N次方差分析.我是否需要像下面这样指定每个因素，或者为此有一个简写形式?
As far as I know (may be I am wrong) that this performs 1-way ANOVA at each factor alone. For some reasons, I need to perform N-way ANOVA between all the 100 groups i.e from x1 to x100. Do I need to specify each factor like the following or there is a shorthand notation for this?
model.lm<-lm(y~x1*x2*x3....,x100, data=data)
anova(model.lm)
推荐答案
您可以使用update.formula和~(.)^n表示法.
例如，对于一个模型，该模型包括来自4个变量a，b，c和d的三向交互作用
Eg for a model including 3-way interactions from 4 variables a, b, c and d
update(~a+b+c+d, ~(.)^3)
## ~a + b + c + d + a:b + a:c + a:d + b:c + b:d + c:d + a:b:c + a:b:d + a:c:d + b:c:d
因此，对于您要适合100次互动的示例，我建议考虑使用一种更合适的模型(尤其是在该考虑的时候).
So for your example where you want to fit 100-way interactions, I would suggest thinking of a more appropriate model (especially if it is time you are accounting for here).
如果您决定继续使用基本的ANOVA方法，则可以执行以下操作(并等待R因大数据/不合适的模型而导致内存问题而崩溃.)
If you decide to continue with the basic ANOVA approach you could do something like this (and wait for R to crash due having memory issues due to your large data / inappropriate model.)
xvars <- paste0('x',1:100)
oneway <- reformulate(termlabels= xvars, response = 'y')
horribleformula <- update(oneway, . ~ (.)^100)
horriblemodel <- lm(horribleformula, data=data)
或者((感谢@Dason帮忙)
Or (thanks to @Dason for picking this up)
stillhorrible <- lm(y ~ .^100, data = data)
这篇关于R中的N向ANOVA的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！