Comparing two matrix in Matlab

1 vue (au cours des 30 derniers jours)
john vattic
john vattic le 26 Juin 2014
Commenté : Anuj Kumar le 24 Déc 2020
I have two matrices x and y, both are results from different algorithms/routines that are supposed to calculate the same result. While I know that the isequal() would check if x and y are the same matrix, the entries in those matrices would not be exactly the same (i.e. some entries may be with 5% off in worst case scenario). In this scenario, what would be the best method of comparing them to see if they are close enough to be considered the same result? Thanks in advance for the advices.
  1 commentaire
Anuj Kumar
Anuj Kumar le 24 Déc 2020
you can try this,assumming x and y has same dimensions
len=lenght(x);
tolerance=x-y;
check=tolerance<0.001;
sum_Check=sum(check)
if sum_Check==len %then results can be accepted

Connectez-vous pour commenter.

Réponses (2)

Titus Edelhofer
Titus Edelhofer le 26 Juin 2014
Hi,
this question is rather tricky to answer in general. A simple approach is to compute the relative error. That works fine as long as there are no zeros in it, i.e.,
absError = abs(A-B);
relError = max(absError(:) ./ abs(A(:)));
This also assumes, that A and B are at least comparable in size, so it does not make a difference if you devide by A or B.
Taking care of all eventualities (NaN, inf, zeros ...) makes it more difficult to answer.
Titus
  1 commentaire
john vattic
john vattic le 26 Juin 2014
i got inf as result of relError

Connectez-vous pour commenter.


Anastasios
Anastasios le 26 Juin 2014
I assume you have some floating point variables, in order to have this 5% error. Try to define a function to check if they are relatively equal, such as
function test = isequalRel(x,y,tol);
test = ( abs(x-y) <= ( tol*max(abs(x),abs(y)) + eps) );
if (! test)
printf("fail x=%e y=%e tol=%e\n", x, y, tol);
end
So basically you do the following :
|a-b| < tol*max(|a|,|b|) + tol_floor
  3 commentaires
Anastasios
Anastasios le 26 Juin 2014
Use the following command
abs(A-B) <= ( tol*max(abs(A),abs(B)) + eps)
where the tol = 0.005 (5%) You will get a logical answer if they are within this error of 5% the answer would be 1 otherwise 0.
Hope it helps
vanita
vanita le 6 Août 2014
how to decide tolerance ? is their is any standard?

Connectez-vous pour commenter.

Catégories

En savoir plus sur Creating and Concatenating Matrices dans Help Center et File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by