Title: Mining Web Access Patterns for Prediction and Document Prefetching

Date: Monday, 12 March 2001

Time: 4:00pm - 5:00pm

Venue: Lecture Theater F (Leung Yat Sing Lecture Theater), Academic Concourse (near lift nos. 25/26), HKUST

Abstract:
Web logs provide a vast amount of information for a system to predict users' probable actions in the future. Accurate prediction makes it possible to design better systems for web document caching, prefetching and adaptation of web-user interfaces. In this talk, I'll present our work in learning and refining predictive models for encoding the path profiles of users from realistic web-log files. These models are used to predict users' future requests within a finite time window. Based on the learned models, we construct a new caching and prefetching policy that produces better performance results than traditional policies. I will also discuss our work in using similarly constructed models for building adaptive web-user interfaces.

Biography:
Qiang Yang's interests include intelligent agents, web-data mining and information retrieval and applied artificial intelligence (case based reasoning, machine learning, knowledge-based software engineering and planning). He held an endowed chair position in the School of Computing Science in Intelligent Software Systems, at Simon Fraser University (SFU) BC, Canada. He received his PhD in computer science at University of Maryland College Park in 1989. Prior to joining Simon Fraser University, he was a faculty member at the University of Waterloo. He is currently a professor at Simon Fraser University.

Yang has published two books and many articles in major conferences and journals. He has been a program committee member of the past AAAI and IJCAI Conferences. He was a Conference Co-Chair for the 2000 Canadian AI Conference, and the Conference Chair for the International Case-Based Reasoning Conference in 2001.