Effacer les filtres
Effacer les filtres

How to make the inverse operation of a " \ " operator ?

3 vues (au cours des 30 derniers jours)
Carlos A. Morales Rergis
Carlos A. Morales Rergis le 1 Août 2013
Lets have a square matrix A, and a rectangular transformation X, such that B has a lower order than A and one has the following expresion:
B=(X\A)*X;
If experimentally somehow One alreaddy has the datta from both B and X, how to solve the expresion for A without using pinv(X)? (wich actually is not accurrate)
  2 commentaires
Jan
Jan le 2 Août 2013
Why do you assume that the PINV method is not "accurate"?
Carlos A. Morales Rergis
Carlos A. Morales Rergis le 2 Août 2013
well actually I've tried, knowing A and X, and finding an experimental approximation to B, the fact is that rebuilding A with that datta and assuming that there are no messuarement mistakes the reconstruct datta does not match in any sense with its original set.
Anyway... perhaps in order to have a both sides transform use a sentence like B=pinv(X)*A*X;
the fact is that because of the size of the matrixes matlab itself recomends instaed of pinv(X) the use of " \ " operator...
what i mean is that I cannot mix both of the "pinv" and " \ " am i right??

Connectez-vous pour commenter.

Réponse acceptée

Roger Stafford
Roger Stafford le 2 Août 2013
In general there will not be enough information to deduce A from B and X. Consider the simple case where B is 1 x 1, X is 2 x 1, and the unknown A is 2 x 2. The matrix equality Y = X\A is overdetermined and has a least squares solution which yields two equations, and the matrix equality Y*X = B will yield only one further equation. This gives three equations and six unknowns for Y and A, which are certainly not sufficient to determine the four unknown elements of A.

Plus de réponses (0)

Catégories

En savoir plus sur General Applications dans Help Center et File Exchange

Produits

Community Treasure Hunt

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

Start Hunting!

Translated by