R - 使用组内的回归函数

Question

Suppose I have a dataframe df with three variables df$x , df$y , df$z , and there is a grouping variable df$g .

Usually, to compute a function WITHIN each group, I do the following

df$new<-unlist(tapply(df$x,df$g,FUN=myfunc))

Now suppose I want to generate residuals from regression of x on y and z WITHIN each value of group g , how do I implement it?

More specifically, without using groups, I would have done

df$new<-resid(lm(df$x ~ df$y + df$z, na.action, na.exclude))

One solution to carry out the previous operation WITHIN groups is to use a loop over unique elements of `df$g', but it would be great if there is any vectorized solution.

Answer 1

library(plyr)
ddply(mydata,.(g),transform, new=resid(lm(x ~ y + z, na.action, na.exclude)))

Test using mtcars data:

mydata<-mtcars

myres<-ddply(mydata,.(carb),transform, new=resid(lm(mpg ~ disp + hp))) # g=carb, x=mpg,y=disp,z=hp
> head(myres)
   mpg cyl  disp  hp drat    wt  qsec vs am gear carb         new
1 22.8   4 108.0  93 3.85 2.320 18.61  1  1    4    1  0.20604566
2 21.4   6 258.0 110 3.08 3.215 19.44  1  0    3    1  2.03023747
3 18.1   6 225.0 105 2.76 3.460 20.22  1  0    3    1 -2.39754247
4 32.4   4  78.7  66 4.08 2.200 19.47  1  1    4    1  1.31212635
5 33.9   4  71.1  65 4.22 1.835 19.90  1  1    4    1  2.60271481
6 21.5   4 120.1  97 3.70 2.465 20.01  1  0    3    1  0.03913515

Answer 2

In data.table you can use by

library(data.table)
DT <- data.table(df)


DT[, new := resid(lm(x ~ y + z, na.action, na.exclude)), by = g]