Finding the centroid?

[Barry Earsman]

Hi folks,

Great library. I'm thoroughly enjoying exploring the world of Semantic
Vectors! Thanks for this valuable contribution.

One thing I find apparently missing. Although I seem to recall reading that
finding "centroids" is simple, there doesn't seem to be a function for
calculating that. I.e, given a set of vectors, is there a function that will
calculate a new vector that represents the point that is closest to all of
them?

Intuitively, this would seem to be the mean of the values for each
dimension, but I'm not sure.

Thanks
Barry

[Christian Prokopp]

Hi Barry,

Sum all vectors in your group and normalize the resulting vector to
calculate the centroid. This assumes that SV uses normalized vectors.

Cheers,
Christian

Post by Barry Earsman
Hi folks,
Great library. I'm thoroughly enjoying exploring the world of Semantic
Vectors! Thanks for this valuable contribution.
One thing I find apparently missing. Although I seem to recall reading that
finding "centroids" is simple, there doesn't seem to be a function for
calculating that. I.e, given a set of vectors, is there a function that will
calculate a new vector that represents the point that is closest to all of
them?
Intuitively, this would seem to be the mean of the values for each
dimension, but I'm not sure.
Thanks
Barry
--
You received this message because you are subscribed to the Google Groups
"Semantic Vectors" group.
To unsubscribe from this group, send email to
.
For more options, visit this group at
http://groups.google.com/group/semanticvectors?hl=en.