Research On Query Expansion Of Patent Domain Information Based On Semanitics And Contexts

Patent information records the wisdom and achievements of the human. It is oneof the most largest information resources in the world and also one of the mostimportant strategic resources. It relates to the latest information and the most activeinnovation in all the fields of technology and has reference value not only in theinnovation of science and technology but also in the strategic decision. However, thequantity of the patents information is huge, the information retrieval on the patents isbased on keywords searching. As a result of language diversity and fuzziness, itbecomes very difficult to realize comprehensive and accurate searching in patentsinformation. While, how to effectively search the patent information that we need toimprove the efficiency of patent information query becomes the core task of thispaper.This paper uses the method of combining semantics and context to achieve theconstruction of query expansion expression in patent information domain, thus,improve the efficiency of patent information retrieval. After the full study of thepatent information, the application status of the technology of text mining andontology, we propose the research contents of this paper. First, it extracts patentinformation through information extraction system developed by our college, thenuses business intelligence services and text mining techniques to process the texts,combining the structural characteristics of patent information and IPC （PatentInternational Classification）number. Thus, it completes the building of patents corpus.Second, it obtains domain ontology concepts from the corpus and mines therelationships between concepts, using text mining technology. Then, it uses theontology editing tool Protege4.1to edit and store the ontology information and usesthe reasoning machine based on ontology to finish the reasoning and validation of theinformation. Finally, it develops the patent information query expansion system usingontology query expansion algorithm to realize the query expansion. Then, it sets thetests to compare the results of search between keywords searching and queryexpansion, which is to validate that the query expansion method based on semanticsand context of the patent information is effective.