2

嗨：我有 10 个因变量和三个自变量的数据集。我想对每个单独的因变量做一个 OLS，依次添加每个自变量。所以，使用 tidyr 我得到了一些模型。

我想对每个因变量进行方差分析，以相互测试嵌套模型。但是在一堆模型的背景下我该怎么做呢？

library(tidyr)
#Fake data, two dependent variables, two independent variables
dv1<-rnorm(100, mean=10, sd=2)
dv2<-rnorm(100, mean=5, sd=3)
iv1<-rnorm(100, mean=100, sd=20)
iv2<-rnorm(100, mean=100, sd=20)
#make a data rame
df<-data.frame(dv1, dv2, iv1, iv2)
#save in model.list
model.list<-df %>% 
  #gather the dependent variables
  gather(variable, score, -iv1, -iv2) %>% 
  #group by dependent variables
  group_by(variable) %>% 
  #run the model against each of the dependent variables, sequentially, first with iv1 then with iv1 and iv2
  do(model1=lm(score~iv1, data=.), model2=lm(score~iv1+iv2, data=.))
4

0 回答 0