USING THE STRUCTURAL CONTENT OF DOCUMENTS TO AUTOMATICALLY GENERATE QUALITY METADATA

摘要

Giving search engines access to high quality document metadata is crucial for efficient document retrieval efforts on the Internet and on corporate Intranets. Presence of such metadata is currently sparsely present. This paper presents how the structural content of document files can be used for Automatic Metadata Generation (AMG) efforts, basing efforts directly on the documents' content (code) and enabling effective usage of combinations of AMG algorithms for additional harvesting and extraction efforts. This enables usage of AMG efforts to generate high quality metadata in terms of syntax, semantics and pragmatics, from non-homogenous data sources in terms of visual characteristics and language of their intellectual content.