Classification of a matrix to 0 and 1 matrix

Hello everyone,

I want matrix A to be like matrix B

ID, age and sex groups are repeated in matrix A. Matrix B is classified age based on the sex group with counting the value from matrix A. if any value in any group was repeated more than 1, then total will appear in matrix B. For example, in matrix A: ID=5, Age group=2, Sex group=2--->Then in matrix B: the value (4,5) is equal by 2

Answer 1

[Roger Stafford]

Assuming A and B are numerical arrays arranged as shown in your diagram,

   B = accumarray([2*A(:,2)+A(:,3)-2,A(:,1)],1,[2*max(A(:,2)),max(A(:,1))]);

[Moe]

Thanks Roger. I'm wondering if we can have matrix B same as follow picture:

[Roger Stafford]

For the second version, just interchange the first and second columns in the formula:

 B = accumarray([2*A(:,1)+A(:,3)-2,A(:,2)],1,[2*max(A(:,1)),max(A(:,2))]);

Answer 2

[Jos (10584)]

A = [1,7,1 ; 2,2,1 ; 2,4,2 ; 3,13,2 ; 3,11,2 ; 4,6,2 ; 5,2,2 ; 5,2,2 ; 5,9,1 ; 6,7,1 ; 6,10,2 ; 7,8,1 ; 7,6,2 ; 7,6,2 ; 7,6,1 ; 7,1,1 ; 7,12,2];

% If A is as above:
nID = 7 ;
nAge = 13 ;
nSex = 2 ;
B = reshape(accumarray(A(:,[3 2 1]),1,[nSex nAge nID]) ,[],nID)

[Joseph Cheng]

take a bit of time to learn what Jos did and (the long way i had above ) and you'll see all you needed to do is

B = reshape(accumarray(A(:,[3 1 2]),1,[nSex nID nAge ]) ,[],nAge)

which is switch out the rows and column call outs that Jos initially wrote out.

[Jos (10584)]

Thanks for the correction Joseph. I didn't bother to check that properly.

[Joseph Cheng]

hmmm? there seems to be a missing comment before mine where the person posting asked for it the other way around. so I commented how to rewrite your function. Jos you had it correct the first time corresponding to the question asked.

[Moe]

Thanks Jos & Joseph Cheng.

Answer 3

[Joseph Cheng]

I would first create a matrix Btemp of size max(AgeGroup) by max(ID) by max(SexGroup) full of zeros. then do a loop for each row of A to add Btemp(AgeGroup,ID,SexGroup) with 1; after you loop for each row of A then make your B matrix by stagering both sexgroup

Air coding so pardon any syntax mistakes:

Btemp = zeros(max(A(:,2)),max(A(:,1)),max(A(:,3))); %create all zeros

%add 1 for each instance listed in matrix A.  
for row=1:size(A,1)
   Btemp(A(row,2),A(row,1),A(row,3)) = Btemp(A(row,2),A(row,1),A(row,3))+1;
end

B=zeros(max(AgeGroup)*2,max(ID)); 
%every other row (even and odd) are the sexgroups.  sexgroup1 is 1,3,5....  sexgroup2 is 2,4,6...
B(1:2:end,:) = Btemp(:,:,1);
B(2:2:end,:) = Btemp(:,:,2);

i think that should do it. or at least gives you a good starting point to correct my 5 min code.

[Moe]

Joseph! I couldn't run/edit your code, could you edit it?

A= [1,7,1;2,2,1;2,4,2;3,13,2;3,11,2;4,6,2;5,2,2;5,2,2;5,9,1;6,7,1;6,10,2;7,8,1;7,6,2;7,6,2;7,6,1;7,1,1;7,12,2];

[Joseph Cheng]

i got lazy and all you needed to do was switch out AgeGroup and ID with A(:,2) and A(:,3)