"Patterns in Search: Mining Web Search User Behavior"
Eugene Agichtein, Emory University

12:00 Noon on Thursday, February 22, 2007
TSRB 132

Accurately modelling user interactions in web search is crucial for search engine development, deployment, and maintenance. I will describe our recent progress in modeling the behavior of web search users to predict search result preferences. Our new general framework for describing search user behavior results in higher accuracy than models based on clickthrough information alone. I will also present two applications of our user behavior modeling techniques. First, we demonstrated that we can significantly improve ranking accuracy by incorporating user behavior features. Second, we applied our framework to automatically identify navigational queries and corresponding "best bet" web search results. The talk will be largely based on my recent work at Microsoft Research. I will also briefly overview my current work on text data mining for improving access to information in large text datasets.

Eugene Agichtein is an Assistant Professor in the Mathematics & Computer Science Department at Emory University. Previously, Eugene was a Postdoctoral Researcher in the Text Mining, Search, and Navigation group at Microsoft Research, working on data mining for information retrieval. He received a Ph.D. in Computer Science from Columbia University in 2005, and a B.S. in Engineering from The Cooper Union in 1998. Eugene is a recipient of the "Best Student Paper" award at the ICDE 2003 conference, and the "Best Paper Award" at the SIGMOD 2006 conference. Eugene's research interests are in information access, and data mining and knowledge discovery with emphasis on the web and life sciences domains.