Exploiting Social Networks to Recommend Resources in Social Bookmarking Applications

April 27, 2012 – ,


Social networks such as Facebook or Google+ are connecting people all over the world. Due to these social networks, more and more information about social interactions and connections are made available online. This additional information gained from the social network of users could be used to enhance personalized recommender systems, particularly in social bookmarking systems such as delicious or Mendeley.
As a result of the collaborative tagging of resources in social bookmarking systems, a data structure similar to that of the Web is created. This data structure called a folksonomy is formed when users attach keywords (called tags) to Web resources such as websites or blogs. Various graph-based approaches for ranking resources in a folksonomy exist such as FolkRank that is based on Google’s PageRank. These ranking approaches have shown to perform better when considering additional semantic information. Therefore the information found in social networks could improve the retrieval of relevant resources in a folksonomy.

Aim of Thesis

The goals of this thesis are:

  • Analyze how best to extend graph-based ranking approaches for folksonomies with additional information gained from social networks.
  • Develop a concept to extend a standard ranking approach such as FolkRank and implement this.
  • Evaluate this approach by comparing with other standard ranking approaches for folksonomies.


  • An interest in semantics and information retrieval in Web 2.0
  • Java programming skills
  • The thesis can be written in English or in German

download corresponding tendering

Keywords: recommender systems, social networks, ranking

Research Area(s):

Tutor: ,

Student: Clément Benaych

Completed Theses