Research On Patent Technology-efficacy-application Diagram Building And Application Based On Mapreduce Computational Model

Ernst proposed an innovative use of patent information as a technicalmanagement measure an abstract framework, The patent is a direct output of thecreative process, but also due to the entry into force of an invention must be auditedby the government, it is also a objective criteria, what’s more, It is one of the largestnumber of information resources in the world, but also one of the most importantstrategic resource in the today’s era of knowledge economy. It involves the newest andmost active innovation in all areas of technology; it is very important reference valueto technological innovation and strategic decision-making. However, the patentinformation’s voluminous, and the research of it today is mostly focus on the analysisof the situation of the apply in a particular year, or the life circle of particulartechnology based on the apply date of the patent, or further the semantic network ofabstract based on the text mining technology, However, due to the variety ofexpression and the deep hidden of patent value, it is difficult to make use of itcomprehensive and accurate in decision support. How can a comprehensive andeffective mining hidden patent information technology and information to improvethe correctness of the patent information-based decision-making is the core task ofthis study.The paper will construct the technology effect matrix and technology applicationmatrix based on the patent field information by combining text mining and distributedcomputing and makes full use of the patent abstract to achieve its value. Afterresearching the current situation of technology roadmap based on patent information,the whole frame of the paper is proposed. First of all, construct the patent ontologybased on the sentence structure of the patent information, it’s mainly rely on theanalysis of sentence structure to refinement and split the abstract to make it morestructured and save all parts of it in the csv files. Secondly, we design a matrixco-occurrence algorithm to turn the parts of patent ontology into co-occurrencematrixes. Finally, based on the data source of the third chapter and method of fourthchapter, the paper will build the whole technology effect matrix and technologyapplication matrix to support the experts to make decision.