Question

I want to create different matrices in a loop and then combine (either cbind or rbind) them. But the following codes doesn’t work. Why not? And how to fix it?

dependent = matrix(c(30,184,6,106), 2, 2, byrow=T)
independent = c(160,166)
expected = numeric()
{for(i in 1:length(independent))
a = dependent*independent[i]/sum(independent)
expected = cbind(expected,a)}

This gives:

expected
        [,1]     [,2]
[1,] 15.276074 93.69325
[2,]  3.055215 53.97546

This is the result of only using the final iteration of the for loop. So the result is like only 166 is used, but 160 isn’t.

Answer 1

A few comments:

1. Your for loop brackets are in the wrong place. You have:

   R> {for(i in 1:3) 
   +   cat(i, "\n")
   +   cat(i, "\n")
   + }
   1 
   2 
   3 
   3 
   

   instead you should have:

   R> for(i in 1:3) {
   +   cat(i, "\n")
   +   cat(i, "\n")
   + }
   1 
   1 
   2 
   2 
   3 
   3 
   

   When you construct a for loop and ommit the brackets, only the first line after the for statement is used.

2. You can make your for loop more efficient by saving the result of sum(independent) since that doesn’t change with each iteration, i.e.

     for(i in 1:length(independent)){
         a = dependent*independent[i]
         expected = cbind(expected,a)
      }
      expected  = expected//sum(independent)
   

3. In fact you can vectorise the whole calculation

   y = sapply(independent, '*', dependent)
   matrix(y, ncol=4,nrow=2)/sum(independent)