Social tags are free-form text that people in social networks use to label and organize information elements. Tags can be used to create graphs that show associations between people, tags, and documents. Tags can be used in search processes with Web portals. The search may yield a set of tags--a tag cloud--that displays links to the most popular tags in the graphs. Font size is often used to show the frequency or importance of a tag in the tag cloud.

This short paper explains tags, tag clouds, and tag navigation when using tags in search processes. Mesnage and Carman describe prior research in social tagging, and then describe a Bayesian approach to generating tag clouds. Using the Bayesian approach, they demonstrate how to generate a tag cloud, given a query. They extend their analysis to the social network structure. The authors obtained a dataset from a music-sharing online social network that allows researchers to fetch the friends of any user of the system. They created a prototype application, and explain how they plan to evaluate the ability of the approach to easily reach a desired item.

The paper is not a short tutorial on social tagging and tag navigation, but it does summarize the issues and related work on social tagging. It explains some of the authors’ research, and gives insight into why efficient and scalable tag navigation of multiple features is a hard problem.