Mining evolution of structure of semi-structured web data

251 p.Web mining is a converging research area from several research communities, such as database, information retrieval, machine learning and natural language processing. In the literature, many research efforts have been directed to Web mining. However, we observed that existing Web mining approaches mainly focused on the semi-structured and the massiveness properties of Web data, whereas there is no systematic research that discovers knowledge by mining the dynamic property of Web data. We believe that knowledge hidden behind the dynamic nature of Web data is also important and useful in many applications such as dynamic-conscious XML cache strategy, intelligent Web advertisement placing scheme, and event detection from Web data.DOCTOR OF PHILOSOPHY (SCE