Question

I am trying to come up with a design matrix / dummy variable matrix which is unbalanced (in the context of a simulation project). The unequal observations are currently simulated as follows:

set.seed(1)
n_i <- as.matrix(round(runif(3,min=0, max=5)))
n_i
       [,1]
 [1,]    1
 [2,]    2
 [3,]    3

I would like to create a matrix of the following form based on the simulated sample sizes above:

des <- cbind(c(1,rep(0,5)),c(0,rep(1,2),rep(0,3)), c(rep(0,3),rep(1,3)))

Are there simple ways of doing this that don’t involve complicated for-loops? Needless to add that my real data set runs into 1000s of observations.
Would appreciate any help on this.

Answer 1

Try this one

Trt <- factor(x=rep(x=c(1, 2, 3), times=c(1, 2, 3)), labels=c("Trt1", "Trt2", "Trt3"))
model.matrix(~Trt-1, data=Trt)



  TrtTrt1 TrtTrt2 TrtTrt3
1       1       0       0
2       0       1       0
3       0       1       0
4       0       0       1
5       0       0       1
6       0       0       1