Context-based metrics for evaluating changes to web pages

Abstract

The web provides a lot of fluid information but this information changes, moves, and even disappears over time. Bookmark lists, portals, and paths are collections where the building blocks are web pages, which are susceptible to these changes. A lot of research, both in industry and in academia, focuses on organizing this vast amount of data. In this thesis, I present context-based algorithms for measuring changes to a document. The methods proposed use other documents in a collection as the context for evaluating changes in the web pages. These metrics will be used in maintaining paths as the individual pages in paths change. This approach will enhance the evaluations of change made by the currently existing Path Manager, in the Walden's Paths project that is being developed in the Center for the Study of Digital Libraries at Texas A&M University.