hide
Free keywords:
-
Abstract:
Social-tagging communities offer great potential for smart recommendation and
“socially enhanced” search result ranking. Beyond traditional forms of
collaborative recommendation that are based on the item-user matrix of the
entire community, a specific opportunity of social communities is to reflect
the different degrees of friendships and mutual trust, in addition to the
behavioral similarities among users. This paper presents a framework for
harnessing such social relations for search and recommendation. The framework
is implemented in the SENSE prototype system, and its usefulness is
demonstrated in experiments with an excerpt of the librarything community data.
