| The existence, research, and operation of Semantic Web are based on formalized ontology. Ontology is formalized, explicit description and constraint on the shared concepts and their relations. Since ontology was proposed from 90's in 20 centuries, it has been concerned by more and more domestic and international scholars. However, research on ontology has just begun, and its theory, methods need further development. Especially, in the current time, ontology construction cost much labors, money and time. So the ontology construction with high efficiency becomes the neck of ontology and Semantic Web research. How to construct ontology effectively is an unavoidable problem.This paper focuses on the acquisition of Chinese ontology with discussion and study. Firstly, the ontology and ontology learning are summarized, including the major methods and evaluation standard of ontology learning, and some kind of ontology learning tools that are widely used currently.Secondly, to solve the problems of traditional methods of constructing ontology, the method based on statistic and rule has been proposed in the paper. The general and the two key sub-frameworks of the method are built, and the feasibility analyses are carried out. The key technical problems under the frameworks, such as the acquisition and pre-process of corpus, the extraction of nomenclatures and relations are discussed. What is more, every solution of the problem is described in detail.Thirdly, the method of automatic ontology population based on decision tree is presented in this paper. The major task of automatic ontology population is the conceptual classification of instances. The instances are acquired from the ontology and they are used as training samples to build a decision tree. This group of rules can be used to guide and enrich ontology knowledge.Finally, based on the investigation above, a primary test is carried out, and the evaluation of the method has been made after analyzing the test result carefully. |
PDF Full Text Request