4

我有一个矩阵:

1 2 3  
4 5 6  
7 8 1 

我如何使用 matlab 找到这个:

第 1 行:第 3 行
第 2 行:---
第 3 行:第 1 行

我想为每一行都有行索引,女巫有共同的元素。

4

3 回答 3

4

考虑这个

A =  [1 2 3;       %Matrix A is a bit different from yours for testing
      4 5 6;
      7 8 1;
      1 2 7;
      4 5 6];

[row col] =size(A)            

answers = zeros(row,row);     %matrix of answers,...
                              %(i,j) = 1 if row_i and row_j have an equal element 

for i = 1:row
    for j = i+1:row                       %analysis is performed accounting for
                                          % symmetry constraint
        C = bsxfun(@eq,A(i,:),A(j,:)');   %Tensor comparison
        if( any(C(:)) )                   %If some entry is non-zero you have equal elements
            answers(i,j) = 1;               %output               
        end
    end
end
answers =  answers + answers';              %symmetric

这里的输出是

answers =

  0     0     1     1     0
  0     0     0     0     1
  1     0     0     1     0
  1     0     1     0     0
  0     1     0     0     0

当然answers矩阵是对称的,因为你的关系是。

于 2012-10-27T08:57:29.330 回答
1

如果您有很多行和/或长行,Acorbe 提出的解决方案可能会很慢。我已经检查过,在大多数情况下,我在下面介绍的两种解决方案应该快得多。如果您的矩阵只包含几个不同的值(相对于矩阵的大小),那么这应该会很快工作:

function rowMatches = find_row_matches(A)
% Returns a cell array with row matches for each row

c = unique(A);
matches = false(size(A,1), numel(c));
for i = 1:numel(c)
   matches(:, i) = any(A == c(i), 2);
end

rowMatches = arrayfun(@(j) ...
    find(any(matches(:, matches(j,:)),2)), 1:size(A,1), 'UniformOutput', false); 

当您的行数较短时,即当行size(A,2)数较小时,另一种选择可能会更快:

function answers = find_answers(A)
% Returns an "answers" matrix like in Acorbe's solution

c = unique(A);
answers = false(size(A,1), size(A,1));
idx = 1:size(A,1);

for i = 1:numel(c)  
    I = any(A == c(i), 2);
    uMatch = idx(I);
    answers(uMatch, uMatch) = true;
    isReady = all(A <= c(i), 2);
    if any(isReady),
        idx(isReady) = [];
        A = A(~isReady,:);
    end
end
于 2012-10-27T11:01:19.090 回答
0

根据您计划对此输出执行的操作,匹配“第 3 行:第 1 行”可能是多余的。

您已经在输出中以“第一行:第 3 行”的形式进行了此匹配

于 2018-03-19T14:14:29.057 回答