【作者】 安静；

【导师】 徐宝祥；

【作者基本信息】 吉林大学 ， 情报学， 2007， 硕士

【摘要】 本文首先系统地介绍了国内外关于本体论知识的不同见解和观点,主要涉及本体论的概念、分类、作用、构造原则、开发方法等及其应用研究,尤其指出了其在情报学中知识表达和知识组织中的应用,目的在于借鉴成功经验。论文概要讲述了知识模型构建的基本知识,介绍了当前主要的知识建模方法,并总结了其各自的优缺点,遂指出当前知识模型构建中存在的问题,于是引入了知识本体的概念,并给出了其建模的优点。鉴于当前知识模型理论研究的不足和本体论语言表示的优越性,论文构建了基于本体论的知识模型框架,并对其三个组件分别进行了规范说明。然后提出了基于本体论的知识建模的步骤和方法,接着指出了本体论的重用和整合以达到知识共享和重用的目的。论文对可租用住房分配进行了实证研究,完全运用了论文前面所论证的理论、归纳的步骤和方法,充分证明了论文提出的基于本体论的知识建模步骤的可行性。最后,论文对全文研究工作进行了总结,并为所取得的研究成果指出进一步的研究方向。更多还原

【Abstract】 The concept of ontology stems from philosophy. Later with the application in artificial intelligence, computer and network areas, its definition has been incorporated into a number of new content. For people and extensive heterogeneous application, it provides a common understanding of the domain knowledge. And for the third generation Internet (Semantic Web), it assures the quality of content-based knowledge acquisition, interoperability and exchange. It has been applied in knowledge engineering, natural language processing, intelligent system integration, object-oriented design, and other fields. It is precisely because of its advantage in knowledge organization and knowledge expression, now in the field of information science, ontological research has now become a major research focus.Ontology is the clear formal specification of a shared conceptual model. In accordance with the degree of dependence on the field, ontology can be classified into four categories : top ontology, domain ontology, task ontology and application ontology. Ontology plays important roles in the following points : (1) Implementation of domain knowledge sharing; (2) implementation of domain knowledge reuse; (3) analysis of domain knowledge, support domain knowledge acquisition; (4) standardization of knowledge. In the purpose of guiding people to construct ontology, many researchers, from practice, have made a lot of standards for constructing ontology. One of the most influential is Gruber’s five rules : clarity and objectivity, completeness, consistency, greatest monotone scalability and minimal commitment. Construction of ontology can reach common understanding and awareness of the information structure, can reuse the professional knowledge and make the assumptions in the field become more clear, can make the professional knowledge spin-off from knowledge management, and can analyze professional knowledge. There are various methods of developing ontology, such as Mike Uschold & King’s "skeleton", Gruninger & Fox’s "evaluation" ,Berneras et al., Methontology, Sensus, "the seven-step method" ,etc.The most essential object of Information Science is information and knowledge. But from the current knowledge application in the enterprise, far from the desired state, there is the issue of knowledge island. It is mainly because there is lack of a unified knowledge model , making it easy for the same description of the knowledge be different and influence the user’s understanding and sharing of knowledge .Knowledge model itself is a tool to explain "knowledge-intensive information-processing task structure". An application knowledge model can provide the specification for the required data and knowledge structure in use. High-quality knowledge model can not only make the function of knowledge system clear and the human-computer interaction convenient, but more importantly, make the knowledge engineer grasp the identification of the important knowledge , analyze the bottleneck of knowledge application, avoid many unnecessary, repetitive knowledge acquisition. So it improves the working efficiency and improves knowledge quantity and quality in knowledge system.Based on the above understanding, it requires us to build a unified knowledge model. And there is always this or that defect in knowledge expression of traditional knowledge modeling method. On one hand, expressed knowledge can not guarantee the uniqueness of the knowledge in the transfer and sharing process; the other hand, the large number of atomic knowledge environment for complex knowledge representation and reasoning may have combinatorial explosion. Therefore, the ability of conflict resolution and the size ,quantity and organization of knowledge atomic, directly decides the availability of the knowledge representation method. Ontology as the knowledge expression method is guaranteed in the process of transfer and sharing of knowledge to understand the uniqueness and accuracy, etc. Therefore, the introduction of the theory of ontology to knowledge model building catches up with the development of the society. These as the theoretical basis , this paper develops the theory and methods of ontology-based knowledge model building. According to the classification of knowledge areas, this paper constructs the component framework of the ontological knowledge model, which consists of three major components : domain knowledge, reasoning knowledge and task knowledge. All components of the knowledge model need to be built basely on knowledge ontology. Combined the three components of the knowledge model with the ontology original language, using BNF(Backus-Naur Form) paradigm, this paper gives a unified description and expression of the knowledge model .This paper summarizes the methods and steps of ontology-based knowledge modeling, including a total of six steps. Its construction process is a cycle of the loop. If the knowledge model works, it further needs to improve the knowledge base; Instead, it returns to Step III, construction of the initial field model, the cycle repeated until perfect. Step IV is expressed by ontology, so it achieves the knowledge sharing and reuse purposes by the use of ontology reuse and integration. It improves the feasibility of the ontology-based knowledge modeling steps through the practice of hired housing distribution process .However, the work is still not completed, there are many issues that deserves further study. Knowledge modeling process is a complicated and constantly cycle process. In practice , the knowledge modeling steps needs further examination. With the constant updating of knowledge, the modeling tools are required increasingly high. Further research will be the validation for the knowledge model completeness, and the specific application-oriented knowledge model with specific application environment, such as TCM diagnosis, agricultural crops.The establishment of ontology must have the participation of experts, which is relatively more strict and difficult. One of the major shortcomings of ontology is needing the participation of experts. How to acquire ontology through the means of knowledge mining is also the focus of future research. With the development of ontology technology and the update of the corresponding support tools, I believe the construction of knowledge model based on ontology will have broader prospects for development.更多还原