How to randomise numbers in a vector?

Dear all,

Suppose I have this vector x= [1;2;3;4];

How can I randomise it? (i.e. create different combinations of 1, 2, 3 and 4)

Thank you very much in advance, Bianca

Answer 1

[Jan]

See doc randperm:

x = [1;2;3;4]
y = x(randperm(numel(x)))

If this is time-critical, use FEX: Shuffle .

[Bianca Elena Ivanof]

thank you!

[Bianca Elena Ivanof]

apologies for bothering you again.

what if I have x=[1 1; 1 2; 1 3; 1 4] and I want to randomise x(:,2)? Using x(randperm(numel(x(:,2)))) returns me a 4-column matrix of 1's, for example, whereas I would like something like

x= [1 1; 1 2; 1 3; 1 4]
y = randomisation function(x(:,2))
y now looks like this [1 2; 1 1; 1 4; 1 3];

thank you very much in advance.

[Jan]

x(:, 2) = Shuffle(x(:, 2))

Or

xx = x(:, 2);
x(:, 2) = xx(randperm(numel(xx)));

[Bianca Elena Ivanof]

Dear Jan,

Thank you very much for your help. The second option works wonders! I am trying to randomise the subconditions of the main conditions of my experiment - in order for me to do so, there is one last step I need to figure out. I understand if this post is becoming a bit too lengthy for you but, if not, could you please be so kind as to help me understand one more thing? If yes, please see the file attached.

counterbalancing(:,1)= participant ID
counterbalancing(:,2)= speed levels (1, 2, 3) 
counterbalancing(:,3)= conditions (1, 2, 3, 4) per speed level

What I am trying to do is, for every speed level (column 2) per participant (column 1), to randomise the 4 speed level subconditions (column 3) - i.e. to randomise the first 4 rows containing 1, 2, 3, 4 and then the next 4 rows containing 1, 2, 3, 4 and so on and so forth, until the end of the matrix. I very vaguely remember there is a simple of doing computations on every n rows at a time, which doesn't require for loops or anything of the sort.

[Jan]

Y  = reshape(X(:, 3), 4, [])
YS = Shuffle(Y, 1);  % Mix columns
X(:, 3) = YS(:);

Answer 2

[Andrei Bobrov]

Hi Elena!

One way:

x= [1 1; 1 2; 1 3; 1 4]
[~,ii] = sort(rand(size(x,1),1));
out = x(ii,:);

or just

out = x(randperm(size(x,1)),:);

[Jan]

randperm uses the Fisher-Yates shuffle now (as FEX: Shuffle), which is more accurate than SORT(RAND). The later is a stable sort, so if two elements replied by RAND are equal (unlikely, but not impossible) the sorting order is not random. If you e.g. have to shuffle a vector of 2^52 elements, uuhm, well... Who cares.

[Bianca Elena Ivanof]

dear andrei,

thank you for your help.