An Intelligent Approach for Constructing Domain Ontology Using Art2 Neural Network and C-Value Method

Research on semantic webs has become increasingly widespread in the computer science community. The core technology of a semantic web is an artefact called an ontology. The major problem in constructing an ontology is the long period of time required. Another problem is the large number of possible meanings for the knowledge in the ontology. To overcome these problems, one approach is developing ontology learning methods and automating ontology construction process. In this article a novel automated method for ontology learning is proposed. First, domain-related documents were collected. Secondly, the C-value method was implemented for extracting meaningful terms from documents. Then, an ART neural network was used to cluster documents, and terms' weight was calculated by TF–IDF method in order to find candidate keyword for each cluster. Next, the Bayesian network and lexico-syntactic patterns were applied to construct the initial ontology. Finally, the proposed ontology was evaluated by expert's views. The primary results show that the proposed ontology learning method has higher precision than similar studies. Key-Words: Ontology; ART Neural Network; Term Frequency–Inverse Document Frequency (TF-IDF); C-value Method; Bayesian network; Lexico-Syntactic Patterns.