How to find percentage of Similarity between two logical matrices

Question

julian gaviria 2023-1-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1900540-how-to-find-percentage-of-similarity-between-two-logical-matrices

评论： julian gaviria 2023-1-25

The following code calculates the percentage of similarity between the attached logical matrices "c1" and "c2":

row = 168;
col = 390;
% Counting equal elements
equal = c1==c2;
% Number of equal elements
equalCount = sum(equal(:));
similarityPercentage= (equalCount / (row * col) ) * 100;
% Display similarityPercentage
disp(similarityPercentage);

How can I compare the same percentage of similarity only considering the matrix places with "1" values. To note, it's not about counting the number of ones. The aim is to see in which percentage the "1" values appear in both metrices in terms of their number (i.e., quantity) and their position in the respective matrices (i.e., row and col).

P.D., I also calculate the mse values for cross validation.

mse_val=mse(c1,c2);

Many thanks in advance

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

the cyclist 2023-1-25

编辑：the cyclist 2023-1-25

在 MATLAB Online 中打开

FYI, you can do your original calculation with this one-liner:

similarityPercentage = mean(c1==c2,"all")*100

I don't completely understand your question, though. Maybe you could use a smaller example, like

c1 = [1 1 1;
      0 1 1;
      0 0 1];
c2 = [1 0 0;
      1 1 0;
      1 1 0];

and tell us exactly what you expect the output to be?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2023-1-25

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1900540-how-to-find-percentage-of-similarity-between-two-logical-matrices#answer_1156170

在 MATLAB Online 中打开

Wouldn't it just be

row = 168;
col = 390;
% Counting equal elements
equal = c1==c2;
% Number of equal elements
equalCount = sum(equal(:));
% Count the number of 1's total 
oneCount = sum(c1 | c2, 'all')
% Compute the percentage of both having 1 as a fraction of either having 1.
similarityPercentage = (equalCount / oneCount) * 100;
% Display similarityPercentage
disp(similarityPercentage);