Faster Rank Calculations and Indexing - For Loop or Another Way?

Question

Shawn Cooper 2020-6-21

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/552460-faster-rank-calculations-and-indexing-for-loop-or-another-way

评论： Shawn Cooper 2020-6-23

Hello, I'm using the following code to find the rank of all nxn matrices with matrix entries [0 1] and plot them in a histogram.

N = 4;
n2 = N^2;
x = repmat({[0, 1]}, 1, n2);
M = reshape(combvec(x{:}), 4, 4, []);
r = [];
for i = [1:length(M)]
    r = [r rank(M(:,:,i))];
end
hist(r)

As additional possble matrix entries are added ([0 1 2], for example), the number of possible matrices increases exponentially and the speed of the rank() function decreases.

Is there a faster way to do this?

2 个评论
显示无隐藏无

Walter Roberson 2020-6-22

Growing the r array dynamically is really killing performance for you.

Shawn Cooper 2020-6-23

Thanks for the feedback on what the issue was. I appreciate it!

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2020-6-22

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/552460-faster-rank-calculations-and-indexing-for-loop-or-another-way#answer_455164

在 MATLAB Online 中打开

N = 4;
n2 = N^2;
ents = [0 1 2];
tic
[parts{1:n2}] = ndgrid(ents);
toc
tic
pages = reshape(cell2mat(cellfun(@(C)C(:).', parts(:), 'uniform', 0)), N, N, []);
toc
np = size(pages,3);
ranks = zeros(1, np);
tic
for K = 1 : np
    ranks(K) = rank(pages(:,:,K));
end
toc