Weighing the Usefulness of Social Tags for Content Discovery

• Published in: Digital Libraries: Universal and Ubiquitous Access to Information pp. 51 - 60
• Main Authors: Razikin, Khasfariyati, Goh, Dion Hoe-Lian, Lee, Chei Sian, Chua, Alton Yeow Kuan
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 2008
• Series: Lecture Notes in Computer Science
• Subjects: Content discovery and organization; Effectiveness; Social computing; Social tagging; Text categorization; Web 2.0
• ISBN: 9783540895329; 3540895329
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-540-89533-6_6

A new wave of social computing applications has empowered users to create and share a variety of content. This upsurge of user-generated data involves a paradigm shift in terms of the management, searching and accessing of information. Social tagging is one of these ways. This paper serves as an extension to the existing work done on investigating the effectiveness of tags for content discovery using text categorization techniques. In particular, we explored how different tag weighting schemes affect classifier performance. Six text categorization experiments were conducted using a dataset drawn from del.icio.us. The results suggest that not all tags are useful for content discovery even with different weights associated with them. Content analysis was done to understand the relationships between the use of a tag on a document and the document’s terms. Implications of this research are discussed.