Problem with for loop

I'a having a problem with the loop.I am doing face recognition in ANFIS. I am using 3 statistical features for 1 image.I already train and test the network.Output for the network in 1 image is:

    output=0.0014
           0.2992
          -0.8178

then, i have to subtract the output with each value in the database so that when the answer is [0.00 0.00 0.00]', it's mean i already find the exact match.

The database is:

    database=0.0014   0.5208   -0.0000
             0.2992   0.2358   -0.0003
             -0.8178   0.4429   -0.0077
             0.0018   -0.0000    0.0008
            -0.0000   -0.0000    0.0131
             0.0000    0.0005    0.4585
            -0.0020    0.9925    1.0074
            -0.0001    0.9557    0.9999
            -0.0035    0.6546    0.9882
             1.0399    0.9322    0.5067
             0.6241    0.9999    1.0000
             0.5647    1.0175    0.9999
             0.9990    1.0006
             1.1715    0.9999
             0.9075    0.9778

i am using for loop to do the process:

     [l,m]=size(database);
     for i=1:1:m
       for j=1:1:l
      diff=anfis_eval(j,i)-anfis_eval1(j,1);
         if(diff==0)
          ans=i:
      end
        end
          end

I manage to get the exact match, my problem now is that what if there is no exact

match and I want to find the smallest different.I think i should put the 'else'

condition so that if the answer is not zero than I shoul go to another condition.

           [l,m]=size(database);
           for i=1:1:m
             for j=1:1:l
      diff=anfis_eval(j,i)-anfis_eval1(j,1);
         if(diff==0)
          ans=i:
         else 
            ?????

end end end

Now, I don't know how to find the smallest different.Can somebody help me?

Answer 1

[Jarrod Rivituso]

Forgive me if you have already thought of this, but couldn't you just keep track of the smallest answer found? Something like...

[l,m]=size(database);
smallestDiff = 1e6; %or some other very high value
for i=1:1:m
  for j=1:1:l
    diff=anfis_eval(j,i)-anfis_eval1(j,1);
    if(abs(diff) < smallestDiff)
      result=i;
      smallestDiff = abs(diff);
    end
  end
end

Also, I would avoid using ans as a variable name, since on the command prompt, MATLAB will sometimes use that variable name as the default return argument to function calls.