Contextual Metadata for Document Databases

Abstract

Metadata has always been an important means to support accessibility of information in document collections. Metadata can be, for example, bibliographic data manually created for each document at the time of document storage. The indexes created by Web search engines serve as metadata about the content of Web documents. In the semantic Web solutions, ontologies are used to store semantic metadata (Berners-Lee et al., 2001). Attaching a common ontology to a set of heterogeneous document databases may be used to support data integration. Creation of the common ontology requires profound understanding of the concepts used in the databases. It is a demanding task, especially in cases where the content of the documents is written in various natural languages. In this chapter, we propose the use of contextual metadata as another means to add meaning to document collections, and as a way to support data integration. By contextual metadata, we refer to data about the context where documents are created (e.g., data about business processes, organizations involved, and document types). We will restrict our discussion to contextual metadata on the level of collections, leaving metadata about particular document instances out of the discussion. Thus, the contextual metadata can be created, like ontologies, independently of the creation of instances in the databases.