Analysis of tag within online social networks

Abstract

In recent years, tagging systems have been paid increasing attentions from both research communities and system designers. Most popular online social networking sites harness tag for managing and locating contents, for organizing and connecting users, and for recommending and sharing resources. We believe that tag acts like bridge between people and resources. Research on tag and tagging behavior will provide us insight about resource space and user activities on the Internet. In this paper, we present a two-level analysis of the tagging system of Del.icio.us. The results from both two levels confirm each other. In network level, we connect tags by users collaborative tagging to form a social network of tags. By investigating its network feature, we find phenomena of small world and scale-free network. We also discover that the links within this network have relatively strong semantic relatedness. In individual level, users' tagging behaviors and patterns are observed by visualizing their bookmarking history on Del.icio.us. Besides, we study the linked users by their tags and find that users within a subscription network share more common interests than random pairs of users. During the analysis, we also discuss the implications of the findings for the design of tag-based system.