array muddle problem

Hi,

I have an array composed of pairs of variables,

X=[a1 a2 b1 b2 c1 c2..]

and I want to modify each PAIR by a small random percentage. My percentage modifier is

percent=linspace(0.95,1.05,100);
rand_percent=randperm(length(percent));
modifier=percent(rand_percent(1:length(X)));

Since 'modifier' is the same length as 'X', if I multiply them together I modify every element of 'X'. However, I need to modify each PAIR by the same percentage, ie

X=[0.99a1 0.99a2 1.02b1 1.02b2 0.95c1 0.95c2..]

I'm getting lost in the arrays on this one, any suggestions?

Answer 1

[Laura Proctor]

Here's a method that works for what you want to do. If X is length n x 1, create the modifier such that it is only n/2 x 1. Then, multiply the modifier by [ 1 1 ] a 1 x 2 array. This results in an n/2 x 2 array. Then, reshape the transpose to create an n x 1 array which will have each element repeated once. Here's the code:

modifier = percent(rand_percent(1:length(X)/2));  
modifier = modifier * [ 1 1 ];
modifier = reshape(modifier',[],1);

Then, the final result can be multiplied element-wise by X:

Y = X.*modifier;

Answer 2

[Paulo Silva]

Laura was faster than me and her code is simple, here's my version

X=[1 2 3 4 5 6]; %just modify X, the code should handle more pairs 
lhX=length(X)/2;
percent=linspace(0.95,1.05,lhX);
rand_percent=randperm(length(percent));
modifier=percent(rand_percent(1:lhX)); 
modifier1=repmat(modifier,2,1); %copy modifier to another line
XX=reshape(X,2,lhX); %put pairs in the same column
XXX=modifier1.*XX; %do the calculation
Xnew=reshape(XXX,1,numel(XXX)); %put all in one line
%Xnew %uncoment this line to see the result