Analyzing and Improving the Performance of Tag-based Systems in the Blogosphere
Tags have recently become very popular as a means for annotating and indexing online content such as blogs, due to their ease of use, as well as the ability to share tags between users. We have been studying the potential advantages and weaknesses of tags as a mechanism for organizing and retrieving content online. We will present a brief overview of the necessary machine learning and information retrieval techniques, and then describe results characteizing the sorts of tasks for which tags are well suited. We will then describe techniques for automatically annotating documents and an algorithm for automatically inducing relationships between tags as a first step to creating a more expressing tagging scheme. We will end with a discussion of current work on automatically recommending tags and visualizing related tags.
Chris Brooks is an Assistant Professor of Computer Science at the University of San Francisco, where he teaches courses on Artificial Intelligence, Distributed Systems, and Computers and Society. He received an M.S. in Computer Science from San Francisco State University in 1997, and a Ph.D. in Computer Science from the University of Michigan in 2002. His research interests include multiagent systems, peer-to-peer systems, machine learning, intelligently dealing with Web data, electronic commerce, the social impact of technology and using technology to address issues of social justice.