12-04-2012, 04:33 PM
Web Page Recommender System based on Folksonomy Mining
for ITNG ’06 Submissions
[Niwa]MiningFolksonomyForWebRecommendation.pdf (Size: 321.81 KB / Downloads: 39)
Introduction
As the scale of the Internet are getting larger and larger
in recent years, we are forced to spend much time to select
necessary information from large amount of web pages created
every day. To solve this problem, many web page recommender
systems are constructed which automatically selects
and recommends web pages suitable for user’s favor.
Though various kinds of Web Pages have been constructed,
there are many points to be improved in them.
Folksonomy and Social Bookmark
Recently Folksonomy and Social Bookmark are getting
popular and spreading widely. Folksonomy is one of the
components ofWeb 2.0 which is famous for SemanticWeb.
Social Bookmark is a web service using Folksonomy.
Folksonomy
Folksonomy [5] is a new classification technique which
may take place of past taxonomy. In case of web page classification
using taxonomy, someone constructs the classification
tree at first, and pages are classified based on the tree.
An example of taxonomy tree is Yahoo Directory. On the
other hand, in Folksonomy, end users put keywords called
tags to each page freely and subjectively, based on their
sense of values. Anyone can choose any word as tag, and
can put two or more tags to one page.
Social Bookmark (SBM)
Social Bookmark (SBM) is a kind of web services on
which users can share their bookmarks. Anyone can see
anyone’s bookmark on SBM. The most popular one is
del.icio.us [6], and has been spread rapidly since the latter
2004. Users can put tags to bookmark pages, based on
Folksonomy. Users bookmarks are related dynamically by
tags.
System Architecture
System outline
As described in Chapter 1, most past web recommender
systems use collaborative filtering. In Basic collaborative
filtering, users preference are expressed as sets of products
they purchased. But it doesn t work well if the number
of products(pages) is too large, because in such case it’s
very hard to find similar users. To solve this problem, we
express users web page preference by affinity level between
each user and each tag .
Conclusion
We constructed a new web recommender system which
is not limited to particular web sites, based on large amount
of public bookmark data on SBM. We also utilize Folksonomy
tags to classify web pages and to express userspreferences.
By clustering Folksonomy tags, we can adjust the abstraction
level of userspreferences to the appropriate level.
We also solved the problem oftag redundancy in Folksonomy
by clustering tags.