Qiang Yang (Computer Science, UST)

Title: Data Mining on the Web

Abstract: With the explosive growth of the web, a tremendous amount of web 
and customer log data are becoming available.  By analyzing these data, we 
can gain valuable insight into users' behavior and apply this knowledge to
improve the performance of web systems.  In this talk, I will present an
overview of our work on web log mining and case mining from customer logs.
I will first describe how to discover association-rule based prediction
models from a web log and apply these models for web caching, prefetching
and user interface design.  I will then discuss how to analyze data logs
accumulated at an web based application server to discover typical cases,
and how to use these cases for web-based diagnosis and recommendation.