Research And Application Of Ontology Semiautomatic Construction In Patent Information Retrieval System

As ontology has a good concept hierarchical structure and supports the logic reasoning, it obtained more and more widespread application in the information retrieval domain, and it raised the recall and precision of the information retrieval greatly. Under such research background, this paper conducted a deep research on semiautomatic construction of ontology, in order to construct a domain ontology which can be applied in the patent retrieval system, and to raise the system retrieval efficiency, enhance the interactive with users and realize the discovery of patent information by the application of ontology in system .The main work of this paper is as follows: Firstly, through the investigation and study of domestic and foreign famous patent retrieval systems, we concluded three aspects which the patent retrieval system can still be further consummated: the efficiency of retrieval, the intercommunion with the user, and the discovery of patent information. Based on the three aspects, the target of constructing and applying ontology in the patent information retrieval system was established.Secondly, this paper proposed a method of constructing the patent domain ontology semiautomaticly, which solves difficult problems in the process of ontology construction: acquiring concepts and acquiring relationships between concepts. An ontology semi-automatic construction system in patent retrieval system was designed. The system can be divided into three major functional modules according to the order of ontology construction, was achieved. The three major functional modules are: constructing ontology prototype module, acquiring concepts module, and acquiring relationships module.And in the process of acquiring concepts, this article proposed a method to calculate the weight of one word in a lot of documents.At last, this article designed and realized the patent information retrieval based-on ontology. This paper proposed the design method of the retrieval module, and realized the semantic retrieval based on ontology with Jena,which is a develpment tool of ontology application.Comparing the ontology-based semantic retrieval method with keyword matching searching method, we find that the former's retrieval result is better than the latter in the recall and the precision of retrieval. The results show that the facet of patent information finding has also been improved. The experiments prove that the method of ontology semi-automatic proposed by this paper is feasible, and application of ontology in the retrieval system perfect the system to some extent.