Effacer les filtres
Effacer les filtres

Random dot product tricks

3 vues (au cours des 30 derniers jours)
Jesse
Jesse le 27 Sep 2013
Commenté : Jesse le 29 Sep 2013
Greetings again,
I have two arrays:
x = [1.3, 5.6, 7.8, -6.5, 0.3]
y = [134.2, 8786.98, -22.4, 4553518.6, 0.0019]
Now, I know how to do the dot product
a=dot(x,y)
and the reverse (i.e, the first dot product has the summation from 1 to 5, but the reverse is 5 to 1),
so
xback = fliplr(x);
yback = fliplr(y);
b=dot(x,y)
but what about largest-to-smallest order dot produc? Is there anyway to organize that in MATLAB from the original arrays?
I didn't know if this was with the sort command.
Thanks!

Réponse acceptée

Walter Roberson
Walter Roberson le 27 Sep 2013
Modifié(e) : Walter Roberson le 27 Sep 2013
xy = x.*y;
dotxy = sum(sort(xy(xy>0), 'descend')) + sum(sort(xy(xy<0)));
  1 commentaire
Jesse
Jesse le 29 Sep 2013
Yes - that's what I was looking for. Very tight too!
Thanks Walter!

Connectez-vous pour commenter.

Plus de réponses (1)

Azzi Abdelmalek
Azzi Abdelmalek le 27 Sep 2013
Modifié(e) : Azzi Abdelmalek le 27 Sep 2013
If you sort a or b or both the result will be different.
Maybe you want to do other thing. Clarify what you want to achieve
% dot(a,b) is the same then sum(a.*b)
  3 commentaires
Azzi Abdelmalek
Azzi Abdelmalek le 27 Sep 2013
No this is not clear, illustrate with a numeric example
Jan
Jan le 27 Sep 2013
Modifié(e) : Jan le 27 Sep 2013
What is your purpose of creating the partial sums at first? Do you want to increase the accuracy of the result? Then better use FEX: XSum: XSum(a .* b).

Connectez-vous pour commenter.

Catégories

En savoir plus sur Matrix Indexing 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