How to find similar elements with small difference (+/- 5%)?

Given: I have 2 arrays. For example: A = [97.2 103.4 109.6]; B = [99.0 105.2 111.4];

Want: I want to find nearest common values with difference +/- 5%. Such as:

    D =
          A         B
       -------   -------
        97.2      99.0
       103.4     105.2
       109.6     111.4

Currently done:

A = [97.2; 103.4; 109.6]; 
B = [99.0; 105.2; 111.4]; 
[E,IA,IB] = intersect(round(A.*0.1),round(B.*0.1),'sorted');
D = [A(IA) B(IB)];

Here is the output:

D =
      A         B
   -------   -------
   97.2000   99.0000
  109.6000  105.2000

As you can see I lost 2 values, and last row is also messed up.

Needed: So I'm wondering, if there are any other functions or code I could use to find similar elements with small difference (+/- 5%).

[ACKNOWLEDGMENTS]

Thanks you: Thorsten,Jan Simon,

I will vote for all your answers.

[MATLAB version]

R2014a

Answer 1

[Thorsten]

One way would be to do this in a straight-forward loop:

p = 0.05; % percentage of difference accepted
C = nan(1, numel(A)); 
for i = 1:numel(A)
  [d ind] = min(abs(B - A(i))); 
  if d < abs(p*A(i)), C(i) = B(ind(1)); end
end
D = [A' C'];

[Ivan Shorokhov]

@Thorsten

Thank you for the great solution, it seems to be working fine for this particular case. I will test it larger data-set and will let you know, whether it works or not.

And there is small mistake in the last line. It should be:

D = [A C'];

Answer 2

[Jan]

What about the new function ismembertol ?

[EDITED] And for older Matlab versions: FEX:ismemberf

[Ivan Shorokhov]

Thank you, I will try it out!