How can i make a partition matrix

I tried to make a mxn partition matrix that all her columns sum 1.0, i have this, but is very slow
c=5; k=3;
matrixTemp = zeros(c, k);
randoms = [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9];
idx_r = zeros(1,k);
for i = 1:c
s = 0;
while s ~= 1
idx_r = randperm(9,k);
s = sum(randoms(idx_r));
end
matrixTemp(i,:) = randoms(idx_r);
end
It is posible more easy?
The partition matrix is the same that fuzzy c means.
I need this for example:
c1 | 0.5 0.6 0.1
c2 | 0,2 0.2 0.5
c3 | 0.3 0.2 0.4
----------------------
R | 1.0 1.0 1.0 % sum each column equals to 1.0

 Accepted Answer

If you are not interested in decimals, just add one, the solution is easy:
c=5;
k=3;
matrixTemp=rand(c,k);
matrixTemp=matrixTemp./sum(matrixTemp)
if you want the entire matrix to have only one decimal place :
c=5;
k=3;
while true
matrixTemp=rand(c,k);
matrixTemp=matrixTemp./sum(matrixTemp);
matrixTemp=round(matrixTemp*10)/10;
if prod(prod(matrixTemp))~=0
break;
end
end
disp(matrixTemp)

1 Comment

Thanks a lot, the first example works fine and faster.

Sign in to comment.

More Answers (1)

If you want to list all possible partitions. Not suitable for large k.
randoms = 0.1:0.1:0.9;
%Index of all permutations with repetitions
[k1,k2,k3] = ndgrid(1:length(randoms)) ; %manually type all k;)
allperm = [k1(:) k2(:) k3(:)] ;
% use the idex to set up permutation matrix of random
random_allperm=randoms(allperm);
%Pick the permuations with sum==1
AllPartitions=randoms(allperm((sum(random_allperm,2)==1),:));
%% pick c number of partitions according to your needs
c=5;
disp(AllPartitions(randperm(length(AllPartitions),c),:))

Categories

Find more on Fuzzy Logic Toolbox in Help Center and File Exchange

Products

Release

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!