Return first unique index of values

   %function [mu,sig,p] = cifar_10_bayes_learn1(f, trlabel,N)
  %trlabel=tr_labels(1:100)';
  label=[1:10,2:11,3:12];
  %N=16;
  mu = zeros(10,(N*3));
  %sig = zeros(10,3)  %task1
  sig = zeros((N*3),(N*3),10); %task2
  p = zeros(10,1);
  for i = 1:10
      [row] = (find(label==i));
        %[row] = find(trlabel==i-1)
  for j=1:N
      mu(i,j) = mean(f((row(i,j))));
      mu(i,j+N) = mean(f((row(i,j+N))));
      mu(i,(j+(2*N))) = mean(f((row(i,(j+(2*N))))));
  end
    mu(i,:)=[mu(i,j),mu(i,j+N),mu(i,(j+(2*N)))]  %error in this, suscripted assignment dimension mismatch
  %     sig(:,:,i) = cov(f((row),:)) 
  %       %Sig(:,:,i)=(cov(f(row))+cov(f(row),:).')/2;
  %       p(i) = (length(row) / length(f))
  end
    end

The above code has to return 10 values in row vector,ie, the first index of unique values.The row values should be displayed as 1,2,3,...,10. I wanted to calculate mean,sigma for that particular indexed rows in training data.

[madhan ravi]

Provide the input datas to test

[Guillaume]

We have no idea what the above code is supposed to do since it's not commented. In addition, it's littered with unnecessary brackets making it harder to read (in my opinion), e.g. the line

mu(i,(j+(2*N))) = mean(f((row(i,(j+(2*N))))));

is simply

mu(i, j+2*N) = mean(f(row(i, j+2*N)));

You should explain the problem you're trying to solve in details, rather than giving us uncommented code that may or may not work.