Skip to main content
eScholarship
Open Access Publications from the University of California

School of Information

Recent Work bannerUC Berkeley

Taggers versus Linkers: Comparing Tags and Anchor Text of Web Pages

Abstract

The Web is home to information services that generate vast quantities of content. This presents a tremendous challenge for services that organize this content and facilitate efficient and relevant retrieval. Tagging is one way to organize this information. Tags are short, user-selected keywords and phrases applied to a Web resource, which can serve as concise semantic descriptions of that resource. Hyperlink anchor text appears to serve similar purposes — namely, as user-generated, concise resource descriptions, useful for navigation and search. In this paper, various aspects of tags and anchors are compared using qualitative characterization, textual properties and usage metrics. We aim to validate perceptions of similarity between these two types of metadata, and to assess the possibility of improving current tag recommendation systems using anchor text. The properties of general level and specific tags are studied and observations are made about how user-generated categorization systems could emerge. A window-based method is used to discover when tags and anchor terms pertain to subtopics within a given document, or to the document as a whole.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View