Beta 1


Title Extending the Wikipedia Recommender System - Assessing Expertise of Recommenders
Author Lefevre, Thomas
Supervisor Jensen, Christian D. (Embedded Systems Engineering, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
Institution Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
Thesis level Master's thesis
Year 2009
Abstract The Wikipedia is a web-based encyclopedia, written and edited collaboratively by volunteers[29]. Over the last couple of years the Wikipedia has surged in popularity, it is by now the de facto place on the internet to look up information on a wide array of subjects. Being as open as Wikipedia is, for new contributors, has a few pitfalls which Wikipedia also openly admits. There has been much dispute about the quality of the work these editors do, Wikipedia's premise is that over time the articles on Wikipedia will improve and eventually become complete and stable. The Wikipedia Recommender System) was developed to help users determine the credibility of an article, to allow users to submit ratings and subsequently receive predicted ratings on articles. The WRS is a rating-based collaborative filtering system that employs trust metrics. Collaborative filtering is a ltering technique that employs the input from other users. The use of collaborative filtering has seen a surge of popularity in recent years with such services as http://delicious.com and http://digg.com.The problem with applying a collaborative filtering system on Wikipedia is the breadth of information available on it and expressing trust unilaterally in a user, without recognizing this breadth. The objective of this thesis is to extend the Wikipedia Recommender System in such a way that it can differentiate between users based on their area of expertise. This will help produce meaningful and trustworthy predictions to the user. It is also demonstrated how the proposed trust model works intuitively and how its prediction results differ in a meaningful way from the previous Wikipedia Recommender System.
Imprint Technical University of Denmark (DTU) : Kgs. Lyngby, Denmark
Series IMM-M.Sc.-2009-75
Fulltext
Original PDF Lefevre.pdf (2.49 MB)
Admin Creation date: 2009-12-22    Update date: 2010-08-25    Source: dtu    ID: 254285    Original MXD