Ontology-based Construction Of Hierarchical Categorization System Of Domain Knowledge Base

Domain knowledge is widely used in many fields, including the analysis of text subject and content. Domain knowledge is the basis of text analysis, so it's very urgent to study how to construct domain knowledge base and make it used better.Ontology can give us basic words and their relations to construct the knowledge base, the first step to construct large data base is to design the related Ontology, and the organization of the knowledge base is very important. Ontology-based construction of the domain knowledge base can not only describe the concepts and their relations clearly, but also realize the reusing and sharing of domain knowledge, and it is also effective in maintaining the domain knowledge base. While constructing domain knowledge base, ontology-based hierarchical categorization system of domain knowledge base is our main research topic.In this paper, we study the definition of domain knowledge base and the representation of domain knowledge, and so on. This paper mainly discuss how to construct domain knowledge base based on Ontology, thereby we can construct the Ontology-based hierarchical categorization system of domain knowledge base.In recent years, there are more and more electronic readable texts in internet. And it's urgent to have a good text categorization system to help us collect and manage all kinds of documents. Text categorization can group the documents into the predefined categories automatically. With the rapid development of information and technology, text categorization has a great future value. However, the current text categorization systems have many problems when they are applied.We attempt to study the application of domain knowledge for text categorization based on domain knowledge. We propose two text representation methods using domain knowledge. One is text representation using domain feature concept, the other is text representation using domain feature attribute.The experimental results show that text representation based on domain...