Vectorizing the creation of an (m x n) matrix where each row represents a number

I am trying to create an (m x n) matrix where every row represents a number. For instance, let's say I wanted to represent the following numbers:

y = [1:3]'

y =

       1
       2
       3

Then that could be represented as:

yk = [1 0 0; 0 1 0; 0 0 1]

yk =

       1     0     0
       0     1     0
       0     0     1

Right now I am using the following code:

yk = zeros(m,n);
for i = 1:m
   yk(i,y(i)) = 1;     
end

But I would like to vectorize it to get rid of the FOR loop. Anyone know how to do this?

Answer 1

[the cyclist]

I'm not sure how your matrix "represents" the vector. I feel that you have not presented a general enough rule for what you want.

Do you just want

eye(3)

?

[Jeffrey Osborne]

Apologies about that. Basically the "n" should be what range of possible values the number could be. So if the set of numbers could be 1:5 (only integers), the n = 5 and (for e.g.) 3 would be represented by [0 0 1 0 0]

Answer 2

[Roger Stafford]

Assuming y is a column vector

yk = zeros(m,n);
yk((1:m)’+m*(y-1)) = 1;

[Roger Stafford]

Perhaps the first line should be:

yk = zeros(m,max(y));

Answer 3

[Walter Roberson]

https://www.mathworks.com/help/nnet/ref/ind2vec.html if you have the Neural Network toolbox