Research On Key Technologies Of Ontology Construction Based On WordNet And Its Application In Security Domain

The knowledge and intelligent is development trend of information technology which makes the representation of information and data not only stay on the level of syntactic, but also focus on the level of semantic and pragmatic. As a concept model which can describe information and data on the semantic level, ontology presents a better way to solve these problems. However, ontology constructing work is basically done manually nowadays, though the concepts and relations between concepts are accurate by the kind of methods, the efficiency can not meet the requirement of development speed. So, the method of constructing ontology automatically is required urgently, though the method has higher efficiency, the accuracy of concepts and its relations are poor because of low performance of technology adopted in the method. According to these deficiencies, this paper presents an ontology constructing method based on WordNet, which can extract ontology from WordNet based on initial ontology. The method has better efficiency because of adopting automatic technology, and ontology constructed by the method has better accuracy because of using WordNet in which the concepts and relations are made manually.Further more, from the view of application, the openness of internet leads to the problems of information content, and they are becoming more and more severe. Constructing security domain ontology related to the content security is an efficient way to solve these problems. The generation of security domain ontology is a further requirement and also is a great challenge of ontology construction method given in this paper, moreover, the information of content security has the characteristics of coving a wide rage and having a highly renewal speed, which increases the difficulty of our work. In this paper, ontology construction method based on WordNet is applied to security domain successfully, and a security domain ontology related to information content security is generated. In order to enrich and perfect the security domain ontology, the key technologies of ontology evolution is studied, the Markov-based concept extraction method and taxonomy learning method based MI matrix are given in this paper. Web information service system based on security domain ontology is given in order to retrieve and extract information of content security and store them in local database, and provide to usersIn this paper, the key technologies of ontology construction based on WordNet, security domain ontology generation and its evolution and application are studied, these achievements are described as follows:1. Proposing the ontology constructing method based on WordNet according to the deficiencies of current constructing methods, especially the low constructing efficiency and poor accuracy of ontology content. The constructing plan is described in the paper: Based on WordNet, giving an initial ontology, and then extracting security domain ontology by semantic similarity method etc. from WordNet based on the initial ontology.2. In order to improve the performance of semantic similarity adopted in ontology construction method of this paper, the research approach is to study the parameter firstly, and presents a DN model of information content of concept in WordNet. The model only relates to the self-structure of WordNet, without participating of other resources. The model not only considers the number of child node of concept in the isa classification tree, but also considers the depth of concept in the classification tree, this made the value of IC more accurate, and the experiment data shows DN model more superior to other models.3. The paper also studies the semantic similarity method based on the parameter, and presents an improved model based on the analysis of current methods. The improved model considers the path between two concepts and their IC, and can improve the accuracy of the semantic similarity value between concepts. The comparison between the improved method and other methods shows the improved method can get better performance.4. Semantic similarity method is the foundation of extracting domain concepts and their relations from WordNet, the extracting approach of this paper is based on the improved model, the extracting process is calculating the semantic similarity value between concepts in initial ontology and WordNet, and then extracting concepts related to security domain according to the given threshold. Subsequently, expanding the isa relationship of initial ontology based on the isa adding rules given in this paper, and generating security ontology at last.5. The construction of security domain ontology is a further demand for ontology construction method based on WordNet, in this paper, the method is applied to security domain successfully and a security domain ontology related to information content security is generated.6. The development of information makes new concepts and new relations of information content security appear continuously. So, the security ontology needs enriching and perfecting constantly. The paper presents a Markov-based concept extracting algorithm and a method of taxonomy learning based on mutual information (MI) matrix.7. Finally, the security ontology constructed in the paper is applied to Web information service system, aiming to retrieving and extracting the information of content security, and storing the data in the local database for user to analysis and research.This paper presents a method of constructing ontology based on WordNet and applies it to security domain, the constructing method also suitable for other domains. The key technologies such as semantic similarity algorithm and Markov-based concept extraction method are superior to other methods, and also can be used in natural language understanding, web page classification and other applications. So, the achievements of this paper have great practical value.