Abstract Study of Design Issues on an Automated Semantic Annotation System

The semantic annotation process turns ordinary HTML web pages into machine-understandable semantic web pages. We have proposed a semantic annotation system to automate annotation of web pages by ontologies. In this paper, we present a study of five design issues on this system, which include: (1) the covering of web pages; (2) the general paradigm of annotation process; (3) the measurements of annotation performance; (4) the compatibility to semantic web standards; and (5) the resolution for lack of ontologies. To address these issues, our proposed system intends to automate annotation for data-rich web pages; simply the annotation process and improve the degree of automation by adapting the ontology-based data recognizer; measure to obtain accurate annotations, to run fast, and to be resilient to page layouts through a two-layer annotation process; be compatible to the semantic web standards by using OWL ontologies; and benefit an interactive ontology creation process by automatically choosing relevant ontology components according to an annotation task from an ontology knowledge base. 1