Cosine similarity between two matrices

31 views (last 30 days)
Paola
Paola on 24 May 2022
Commented: Paola on 30 May 2022
Dear all,
I have some vectors 32x1, representing force fields. I use the quiver function to plot and visualize the fields. (see pic attached) Obtaining a quiver showing 4x4 vectors, for a total of 16 vectors.
X = [1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4];
Y = [1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4];
U = W(1:2:31);
V = W(2:2:32);
figure, quiver(X,Y,U',V');
Even if visually they look very similar, I need to calculate a cosine similarity value, between the different vectors.
Checking online I found that this formula:
cosSim = dot(a,b)/(norm(a)*norm(b));
and also the function
D = pdist(X,'cosine').
I wonder if these two functions can be applied to my case and which one is better to obtain just a similarity value for each couple of vector compared.
Thank you!

Accepted Answer

Matt J
Matt J on 24 May 2022
I tend to prefer pdist2,
UV=[U(:),V(:)];
D=1-pdist2(UV,UV,'cosine')
  5 Comments

Sign in to comment.

More Answers (0)

Community Treasure Hunt

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

Start Hunting!

Translated by