You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have been using somoclu for a week now and really enjoying it. I would appreciate it if cosine distance is also implemented, as it is useful for many document matching application.
Thanks,
Mos
The text was updated successfully, but these errors were encountered:
We routinely use Somoclu in text mining and it makes little sense to implement the cosine distance because it is almost identical to the Euclidean, as long as your data instances are normalized. Take two vectors, x and y. Then (d(x,y))^2=sum_k (x_k-y_k)^2 = |x|^2+|y|^2 - 2(x,y). Since |x|^2=|y|^2=1, this is just 2(1-(x,y)), so a separate cosine distance would result in the exact same topology.
Hi,
I have been using somoclu for a week now and really enjoying it. I would appreciate it if cosine distance is also implemented, as it is useful for many document matching application.
Thanks,
Mos
The text was updated successfully, but these errors were encountered: