How to prepare normalized feature matrix in MATLAB ?

Hello all, I am having a feature matrix say of dimension 500 by 4. I want to normalize this feature matrix and obtain normalized feature matrix such that element of is obtained as ----(1)

where indicates element of feature matrix .

My query is how can we obtain normalized feature matrix from equation (1).

Any help in this regard will be highly appreciated.

[charu shree]

I tried with 'for loop' and getting the proper answer. Here are my efforts:

for i = 1:500

for j = 1:4

p1 = min(D(i,:));

p2 = max(D(i,:));

d_ij = D(i,j);

deno = p2-p1;

numer = d_ij-p1;

P(i,j) = numer/deno;

end

end

Is there any other way so that I can avoid 'for loop' ?

Answer 1

[Dyuman Joshi]

%Random data
D = rand(500,4);

%minimum of each row
minval=min(D,[],2);
%maximum of each row
maxval=max(D,[],2);

%Vector approach
P1=(D-minval)./(maxval-minval);


P2=zeros(size(D));
for i = 1:500
    for j = 1:4
        p1 = min(D(i,:));
        p2 = max(D(i,:));
        d_ij = D(i,j);
        deno = p2-p1;
        numer = d_ij-p1;
        P2(i,j) = numer/deno;        
    end
end
%Checking if the outputs are equal or not
isequal(P1,P2)

ans = logical
   1

[charu shree]

Thank you very much sir....