We are interested in content-based web mining as well as access-based web mining and their applications to metasearch and information filtering.
COMBINATION OF RELEVANCE RANKING FOR WEB METASEARCH
- Faculty: Sergio A. Alvarez, Carolina Ruiz.
- Students: Anna Novikov.
Project DescriptionThis project focuses on a key issue in web metasearch engine design, namely the combination of relevance rankings output by search engines in response to user queries. Combination methods originating in the theory of voting and proportional representation, and a method based on geometric transformations were compared. A system incorporating these methods was designed and constructed. The system adapts to user preferences by selecting search engines based on a metaindex that is updated dynamically via passive feedback from the user.