| With the emphasis on patent data and the progress of computer technology, more andmore researchers choose to get the latest information by patent search, and then they makethe information to be contributed to the continuous improvement and innovation ofproducts and technology. Faced with the rapid growth of the number of patents, users havea wider range of choice in the process of patent retrieval, but there has been a more difficultproblem——how to obtain the information they really need quickly and accurately. Andpatent retrieval is different from the traditional information retrieval–higher request to theusers and grasp of the knowledge of the field to query. There are various problems in thecurrent patent search, such as―polysemy‖,―expression meaning incomplete‖,―can‘tunderstand the real needs of users‖,―too many search results‖. It is lack of guidance to theusers in the retrieval process, and users have the increasing demand for patent search. It ishoped that we can allow retrieval system to assist users‘complete access to informationthrough improved technology, which realizes semantic extension and reasoning in the truesense.In view of the above problems, based on the domestic and overseas research statusanalysis of the domain ontology, semantic retrieval and patent retrieval, this paper putforward the core ideas which do the patent semantic retrieval with the help of domainontology, and designs the framework of patent semantic retrieval system structure. Thenthis paper carried on the in-depth research and analysis on the key technologies and toolsinvolved in the patent semantic retrieval system, which provide a good technical foundationfor achieving semantic retrieval system. These key technologies include the construction ofdomain ontology, semantic reasoning, and the tools include ontology buildingtool—protégé, semantic analysis and reasoning tool—Jena, Java-based developmentplatform——Eclipse, tec. Secondly, some key processes of the patent semantic retrievalsystem are described in detail. This paper achieves a semi-automatic construction ofdomain ontology with the help of protégé, and achieves the reasoning and expansion ofdomain ontology with the help of Jena. Finally, this paper did the integration of several keyprocesses, which realized the patent semantic retrieval system based on domain ontology,by customizing some personalized inference rules.In this paper, the implementation of patent semantic retrieval system based on domain ontology solves the current lack of semantic knowledge retrieval problems and enhancesthe recall ratio of patent retrieval, and it guides the users to define a good retrieval needs inthe search process. Besides, with a case study of the new energy automotive industry, thispaper builds the corresponding domain ontology, designs the personalized semanticinference rules, integrates be the key processes,and then realizes the system. The wholeprocess has certain universality, which can provide other domain ontology semanticretrieval with certain reference significance and reference value. In addition, the method ofsemi-automatic construction of domain ontology and the reasoning process of domainontology will provide some theoretical basis and technical support experience for futurelarge-scale application of the semantic retrieval system. |
