Bringing the Social Semantic Web to the Personal Learning Environment
Vlad Posea
Dept. of Computer Science
University Politehnica of Bucharest
vlad.posea@cs.pub.ro
Stefan Trausan-Matu
Dept. of Computer Science
University Politehnica of Bucharest
trausan@cs.pub.ro


Abstract—The paper proposes a method to semantically index
the learning resources accessible to the learner in the social
websites that he uses and bring them together in a Personal
Learning Environment together with features like tag-based
search and recommendations and learner profile generation.
The paper describes the algorithms used, the semantic model
for storing the data and the widgets that provide feedback to
the learner and that can be used in learning systems like
Moodle or Elgg.
Keywords-social search; semantic web; recommendation;
personal learning environment
I. INTRODUCTION
For a large number of people the internet has become a
socializing environment that we now call the Social Web.
Applications like Facebook, MySpaces, Wikipedia, Youtube
feed on the people’s enthusiasm on participating in social
spaces. Millions of users are using the social’s web features
daily, writing on blogs, creating and sharing bookmarks,
videos and images. Some of these users are parts of
communities of practice in the sense defined by
Lave&Wenger [1]. Some are connected through the social
objects they use [2]. In spite of the huge success of these
social websites a strong limitation is the lack of connectivity
between them. The users of a social website can’t usually
connect or share content with users of other social websites.
Social websites begin to acquire relevance also in the
educational context with learners using social media
applications for learning purposes. Some relevant examples
are the Massachusetts Institute of Technology that has a
channel for posting videos on Youtube that has 47,000
subscribers and whose channel page has been viewed for
more than 1 million times (http://www.youtube.com/mit).
Stanford and Berkeley score similar results with more than 1
million views and more than 30000 subscribers. Delicious
(http://delicious.com) is another useful platform for learning
purposes as many bookmarks point towards educational
content. Slideshare (http://www.slideshare.net) is used for
posting presentations and there are many slideshows posted
by university professors having each of them thousands of
views. The main challenge is to identify relevant content for
the learner and facilitate access to this content inside the
learning platform that the learner uses regularly. This paper
is going to present an application that indexes content from
social websites in a semantic repository and delivers search,
recommendation and profile generation to a learning
environment. The learning scenario of using such an
application is also presented.
The paper continues with the presentation of a method
for semantic indexing of the learner’s social network and
with the description of the search, recommendation and
profiling algorithms that are used for providing feedback to
the learner.
Finally we show how we integrated these services in a
learning platform.
II. SCENARIO – USING THE SOCIAL WEB FOR LEARNING
PURPOSES

Most social websites offer the possibility to tag the
resources posted by the users. Delicious allows multiple
users to tag the same resource while Slideshare or Youtube
allow only the uploader (creator) to tag the resource. The
result of this collaborative tagging is called a folksonomy.
No matter the type of content existing on the social website
(images, videos, bookmarks, presentations) they are all
described by the community of users using tags. These are
used for search and classification on the social websites.
A learner using a social website will use tags to classify
his own content and will also indirectly use tags to find
interesting resources and to find users with which he shares
common interests. The learner might establish relations with
these users. The relations might be unidirectional – the
learner is a called a “fan” or “follower” or “subscriber” of a
user and means the learner is interested on what the user
posts on the website or bidirectional – the learner and user
are “friends”. This means they both acknowledge the
existence of a relationship or they both are interested in the
other’s activities on the website. These connections are
valuable because they can be used to monitor new resources
appearing on the network from relevant users and also allow
discovering new valuable connections from the peers of the
user with witch the learner is connected.
The learner has however difficulties when he needs to
use these connections created in the social websites for
learning purposes. He can’t know which users from all the
websites he’s using have posted resources on a very specific
topic. In order to do this he would have to use the search
facilities of every website and afterwards to see if one of his
peers is among the search results. He also can’t know which
resources recommended or created by his peers can solve his
problem. In order to do this he would have to browse the
resources posted by the users found at the previous step
because there’s no way to search through a user’s resources.
In order to solve the learner’s problems we propose the
following approach. The references (URLs) to the learner’s
2010 10th IEEE International Conference on Advanced Learning Technologies
978-0-7695-4055-9/10 $26.00 © 2010 IEEE
DOI 10.1109/ICALT.2010.48
148