Question

我有一个矩阵

My_big_matrix_=rand(21,4)
   My_big_matrix_=[ 
   1.0000    0.8147    0.0357    0.7655
    2.0000    0.9058    0.8491    0.7952
    3.0000    0.1270    0.9340    0.1869
    4.0000    0.9134    0.6787    0.4898
    5.0000    0.6324    0.7577    0.4456
    6.0000    0.0975    0.7431    0.6463
    7.0000    0.2785    0.3922    0.7094
    8.0000    0.5469    0.6555    0.7547
    9.0000    0.9575    0.1712    0.2760
   10.0000    0.9649    0.7060    0.6797
   11.0000    0.1576    0.0318    0.6551
   12.0000    0.9706    0.2769    0.1626
   13.0000    0.9572    0.0462    0.1190
   14.0000    0.4854    0.0971    0.4984
   15.0000    0.8003    0.8235    0.9597
   16.0000    0.1419    0.6948    0.3404
   17.0000    0.4218    0.3171    0.5853
   18.0000    0.9157    0.9502    0.2238
   19.0000    0.7922    0.0344    0.7513
   20.0000    0.9595    0.4387    0.2551
   21.0000    0.6557    0.3816    0.5060];

我希望获得不同的行组合以获得大小（7 * 4）的矩阵

这些组合应该是

Combination_(1,1)={1,4,7,10,13,16,19}
Combination_(1,2)={1,5,7,10,13,16,19}
Combination_(1,3)={1,6,7,10,13,16,19}
Combination_(1,4)={2,4,7,10,13,16,19}
Combination_(1,5)={2,5,7,10,13,16,19}
Combination_(1,6)={2,6,7,10,13,16,19}
Combination_(1,7)={3,4,7,10,13,16,19}
Combination_(1,8)={3,5,7,10,13,16,19}
Combination_(1,9)={3,6,7,10,13,16,19}
Combination_(1,10)={1,4,7,10,13,16,19}
Combination_(1,11)={1,5,8,10,13,16,19}
Combination_(1,12)={1,6,9,10,13,16,19}

EDITED：在我的组合中，三行（1,2,3），（4,5,6），（7,8,9），......（19,20,21）......是组。只会有一名成员，而且不会有两名成员。所以可能的解决方案不会太高。

My_small_matrix_ having size of (7,4)

有什么想法吗？

Answer 1

C = combnk(1:21,7);   % C rows contains combinations
numcomb = size(C, 1);
combined = zeros(numcomb, 7, 3);
for k = 1:numcomb
    combined(numcomb, :, :) = reshape(My_big_matrix(numcomb(k), :), 1, 7, 3);
end

Answer 2

尝试使用以下代码将数据导入单元格数组，其中每个单元格包含来自每个组合的数据 -

%// Input
My_big_matrix_ = rand(21,4)

%// Use allcomb to generate all such combinations. This a MATLAB
%// file-exchange code, available at -
%// http://www.mathworks.in/matlabcentral/fileexchange/10064-allcomb
allcomb_out = allcomb(1:3,4:6,7:9,10:12,13:15,16:18,19:21);

t1 = reshape(allcomb_out',[],1);%//'
t2 = My_big_matrix_(t1,:);

out = mat2cell(t2,7*ones(1,size(allcomb_out,1)),4); %// Desired output

Answer 3

My_big_matrix_ = rand(21,4); %// data
n = size(My_big_matrix_,1);
g = 3; %// size of a group. Assumed to divide n

m = n/g;
combs = dec2base(0:g^m-1, g)-'0'+1;
combs = bsxfun(@plus, combs, 0:g:g*(m-1));

这给出了

combs =
     1     4     7    10    13    16    19
     1     4     7    10    13    16    20
     1     4     7    10    13    16    21
     1     4     7    10    13    17    19
     1     4     7    10    13    17    20
    [...]
     3     6     9    12    15    18    20
     3     6     9    12    15    18    21

然后你的小矩阵是My_big_matrix_(combs(1,:),:)，My_big_matrix_(combs(2,:),:)等。

从matlab矩阵中获取用户定义的可能解决方案

3 个答案: