基于专利文本挖掘的产业关键共性技术识别与应用研究  

Research on Identification and Application of Industrial Key General Technology Based on Patent Text Mining

作　　者：胡凯 谢芬 杨滨瑜 胡新杰 刘汉霞 Hu Kai;Xie Fen;Yang Binyu;Hu Xinjie;Liu Hanxia(School of Economics,South-Central Minzu University,Wuhan 430074,China;College of Economics and Social Development,Nankai University,Tianjin 300071,China;School of Management,University of Science and Technology of China,Hefei 230022,China;College of International Business and Economics,Wuhan Textile University,Wuhan 430202,China)

机构地区：[1]中南民族大学经济学院,湖北武汉430074 [2]南开大学经济与社会发展研究院,天津300071 [3]中国科学技术大学管理学院,安徽合肥230022 [4]武汉纺织大学外经贸学院,湖北武汉430202

出　　处：《科技管理研究》2023年第8期21-31,共11页Science and Technology Management Research

基　　金：国家社会科学基金一般项目“产业关键共性技术研发的财政激励机制优化研究”(19BJL079)。

摘　　要：突破产业关键共性技术掣肘,首先需要科学识别兼具关键性和共性的产业技术,但目前国内关于关键共性技术识别的研究非常有限。遵循专利文献挖掘思路,兼顾主客观标准,运用大样本量数据集,研究提出一种产业关键共性技术识别方法,为政府制定有针对性的创新支持政策提供决策参考。以我国高校(2001—2020年)和上市公司(2006—2020年)的海量专利作为研究样本,通过构建LDA主题模型、计算专利关键性得分和共性得分,以及运用相似度分析将选择结果与工信部《产业关键共性技术发展指南(2017年)》进行相似度匹配,来识别出我国产业关键共性技术。结果显示:无论是高校还是上市公司,属于关键共性技术专利的比例均不高,二者申请专利中的比例分别为14.7%、10.2%,授权专利中的比例分别为13.8%、10.1%,高校专利占比略高于上市公司;而得分位居前列的技术主题如“电路芯片耦合”“能效模型图像”“系统数据模块”等与集成电路、人工智能、大数据等当前广受关注的关键共性技术非常一致,表明所用方法是有效、可取的。To break through the constraints of industrial key generic technology,it is necessary to scientifically identify industrial technologies that are both critical and general.But there is few research on the identification of key generic technology in China.Therefore,according to the idea of patent literature mining,and taking into account the subjective and objective criteria,this paper uses a large sample size dataset to propose a method for identifying industrial key generic technology,in order to provide a reference for the government to formulate targeted innovation support policies.It takes the massive patents of China's universities(2001-2020) and listed companies(2006-2020)as the research samples to identify the industrial key generic technology of China,by constructing the LDA theme model,calculating the critical and generic patent scores,and using the similarity analysis to match the results with the Industrial Key General Technology Development Guide(2017) issued by the Ministry of Industry and Information Technology.The results shows that the proportion of key general technology patent is both not high in universities and listed companies,of which the proportion of patent applications is respectively 14.7% and 10.2%,and the proportion of authorized patents is 13.8% and 10.1%;the proportion of university patents is slightly higher than that of listed companies.The top scoring technical themes such as "circuit chip coupling","energy efficiency model images" and "system data modules" are highly consistent with key generic technology of current interest such as integrated circuits,artificial intelligence and big data,indicating that this method is effective and desirable.

关 键 词：产业关键共性技术 关键性 共性 专利 文本挖掘 技术主题 阈值选择 

分 类 号：N99[文化科学—情报学] G250.252G301