在R中存储和访问模型结果

Question

I have some code that runs a model in a loop. Each iteration of the loop runs a slightly different model and the results are stored in a variable . What is a good way to store these objects so I can access them after the loop terminates ? I thought about something like this:

fit.list <- list(n)
for (i in 1:n) {
    fit <- glm(......)
    fit.list[i] <- fit
}

But then I want to access each model results, for example summary(fit.list[4]) or plot(fit.list[15]) but that doesn't seem to work.

Answer 1

Try

plot(fit.list[[15]])

The single [ function extracts a list with the requested component(s), even if that list if of length 1.

The double [[ function extracts the single stated component and returns it but not in a list; ie you get the component itself not a list containing that component.

Here is an illustration:

> mylist <- list(a = 1, b = "A", c = data.frame(X = 1:5, Y = 6:10))
> str(mylist)
List of 3
 $ a: num 1
 $ b: chr "A"
 $ c:'data.frame':  5 obs. of  2 variables:
  ..$ X: int [1:5] 1 2 3 4 5
  ..$ Y: int [1:5] 6 7 8 9 10
> str(mylist["c"])
List of 1
 $ c:'data.frame':  5 obs. of  2 variables:
  ..$ X: int [1:5] 1 2 3 4 5
  ..$ Y: int [1:5] 6 7 8 9 10
> str(mylist[["c"]])
'data.frame':   5 obs. of  2 variables:
 $ X: int  1 2 3 4 5
 $ Y: int  6 7 8 9 10

Notice the difference in the last two command outputs. str(mylist["c"]) says " List of 1 " whilst str(mylist[["c"]]) says " 'data.frame': ".

With your plot(fit.list[15]) you were asking R to plot a list object not the model contained in that element of the list.

Answer 2

also maybe try

fit.list <- list()
for (i in 1:5) {
counts <- c(18,17,15,20,10,20,25,13,12)
outcome <- gl(3,1,9)
treatment <- gl(3,3)
print(d.AD <- data.frame(treatment, outcome, counts))
glm.D93 <- glm(counts ~ outcome + treatment, family=poisson())
fit.list[[i]] <-glm.D93
}

note the fit.list[[i]] rather then fit.list[i] as you have