Article Preview
Top1. Introduction
The World Wide Web is a massively distributed reservoir of information, but the information does not have well-defined machine-understandable meaning attached to it, prohibiting automated manipulation and reasoning about such information (Ram & Zhao, 2007). The next generation of the World Wide Web, the Semantic Web, is intended to enable more intelligent use of data and information for effective electronic interoperability and collaboration (Horrocks, 2008). A successful Semantic Web, however, depends upon the ability to manage, integrate, and analyze data and is driven by the role of semantics for automated approaches to exploiting Web resources (Berners-Lee, Hendler, & Lassila, 2001). Ontologies, which are at the heart of the Semantic Web, define the concepts and relationships that make global interoperability possible, facilitate sharing and integration (Horrocks, 2008; Leukel & Sugumaran, 2009; Tun & Tojo, 2008) and serve as surrogates for semantics. Ontologies are also useful for digital libraries and personalized information management (Katifori, Halatsis, Lepouras, Vassilakis, & Giannopoulou, 2007; Meng & Chatwin, 2010). Although their need is well-documented, ontology development is often performed manually and is challenging and time-consuming (Ding & Foo, 2002; Farquhar, Fikes, & Rice, 1997). One of the major reasons for this difficulty is finding relevant knowledge sources to use to create ontologies.
The World Wide Web is a great resource of information for almost all imaginable domains. If this information could be properly extracted and organized, it should be possible to effectively use it to create domain ontologies, especially if a process to do so could be automated to some extent (Sánchez & Moreno, 2008). The objectives of this research, therefore, are to:
- ⁃
Develop a methodology for semi-automatically generating domain ontologies by extracting and organizing terms and relationships among those terms using the World Wide Web as a source;
- ⁃
Establish the feasibility of the ontology creation methodology by creating a prototype; and
- ⁃
Assess the performance of the methodology through an empirical analysis.
The contribution of the research is to develop a way to semi-automatically create domain ontologies by using the World Wide Web as a source and integrating web tools. Libraries could be used for the Semantic Web and other applications (e.g., heterogeneous databases, conceptual modeling, and web queries (Horrocks, 2008; Ram & Zhao, 2007)).
Section 2 of this paper examines related research on domain ontologies and their role in the Semantic Web. A six-step ontology creation methodology is presented in Section 3. Section 4 details the implementation of the methodology in a prototype, WebtoOnto. Section 5 evaluates the methodology using an empirical study. A summary and concluding remarks are found in Section 6.