Extract data from a vector in a loop

Hi everyone, First of all, I am a beginner.

I would like to make a function that would do the variance over specified number of samples "n". I need a variance of a signal (which is in a vector "x"). Basically, from a matrix "x", it takes first "n" number of samples, calculates variance and carries on to the next "n" samples. Now, I need those variance calculations in one vector, so I can plot it. That vector would have "length(x)/n" units. I made something like this:

function [a,b] = Variance( x, n )

 a=zeros(1,length(x));

 b=zeros(1,length(x));

 for i=1:length(x)/n

    a=x(i*n:(i+1)*n);

    b=var(a);

 end
 end

This, of course does not work. I would appreciate if you could help out.

Answer 1

[Joseph Cheng]

So lets break it down. in your for loop

for i=1:length(x)/n

    a=x(i*n:(i+1)*n);

    b=var(a);

end

when i=1 you are getting x(1*n:2*n) and missing the the values of x from 1 to n. So you'll want to start with i-1 and go to the i value. example:

    x=1:20;
    for i=1:10
         disp(x((i-1)*2+1:(i)*2))
    end

Since you are assigning the calculations to a and b you should additionally be putting them into the correct array positions (i). so you'll have a(i) = __ ; and b(i) =var(a(i)); that way you get your length(x)/n arrays.

Another method of this is you can use reshape. example:

x=1:20
y = reshape(x,5,4);
b=var(y);

here you see that the reshape formed your buckets of n-samples for you.

~J

[Plastic Soul]

Hi Joseph,

Thanks for a fast reply. I modified the array positions with a(i) and b(i), but now, I am getting error:

In an assignment  A(I) = B, the number of elements in B and I must be the same.

Here is the code:

function [a,b] = Variance( x, n )

a=zeros(1,length(x)/n);
b=zeros(1,length(x)/n);

for i=1:length(x)/n
    a(i)=x(n*(i-1)+1:i*n);
    b(i)=var(a(i));
end

end

[Plastic Soul]

Never mind, I think I did it with the other method:

function [ b,a ] = Var1( x,n )
a=reshape(x,n,length(x)/n);
b=var(a);
end

Seems like its working... Thanks a lot. If you notice anything wrong, let me know pls.