Architecture of WebOntEx: A System For Semi-Automatic Extraction Of Ontologies From Web Pages

摘要

The Internet has been developing rapidly and the amount of information on the Internet is growing tremendously. To manage this huge growing data and to discover valuable information from it, incrementally extracting and managing metadata on the Web is needed. Hence, there is a need to extract conceptual structure (i.e., ontologies) of Web data and building a meta-database. The goal of the WebOntEx (Web Ontology Extraction) project is to extract Web ontologies semi-automatically by analyzing Web pages that are in the same application domain and to convert the ontology to XML DTD (Document Type Definition). The extracted ontologies can be used in various important applications, such as understanding Web information content, querying Web metadata, more intelligent Web searching, and conversion of unstructured/semistructured HTML Web pages to other formats, such as XML. The ontology is considered as a complete schema of the application domain concepts as they are used in the analyzed Web pages. The concepts are classified into entity types, relationships, attributes, and superclass/subclass hierarchies and stored in a relational database to allow them to evolve over time. We utilize machine-learning techniques, in particular inductive logic programming, in the WebOntEx Heuristic Analyzer module. This paper describes the WebOntEx project, and its architecture.