Difference between revisions of "Cosine similarity"

From RecSysWiki
Jump to navigation Jump to search
(new page)
 
Line 1: Line 1:
 
The '''cosine similarity''' measures the similarity between two vectors by the cosine of the angle between them.
 
The '''cosine similarity''' measures the similarity between two vectors by the cosine of the angle between them.
In [[recommender systems]], the cosine similarity is often used in [[kNN]]-based methods to determine the neighborhood of a [[user]] or [[item]],
+
In [[Recommender System|recommender systems]], the cosine similarity is often used in [[kNN]]-based methods to determine the neighborhood of a [[user]] or [[item]],
 
and as a weight to compute predictions using those neighbors.
 
and as a weight to compute predictions using those neighbors.
  
 
== External links ==
 
== External links ==
 
* [[Wikipedia: Cosine similarity]]
 
* [[Wikipedia: Cosine similarity]]

Revision as of 08:58, 22 February 2011

The cosine similarity measures the similarity between two vectors by the cosine of the angle between them. In recommender systems, the cosine similarity is often used in kNN-based methods to determine the neighborhood of a user or item, and as a weight to compute predictions using those neighbors.

External links