摘要
随着我国企业每年专利申请量的不断增多,一些专利申请乱象日渐显现:国家为专利申请制定了扶持与鼓励政策,有的企业为了争取这些政策奖励而胡乱申报或购买对于企业发展与利润提升无积极影响、且与企业主营产品无关的专利。在这种背景下,专利与企业相关性测度研究就成为一个有价值的研究课题。针对当前专利与企业相关性判断依赖领域专家的主观评价现状,提出一种专利与企业相关性的自动测度方法,为传统主观评价方法提供替代思路。采用信息论与深度学习方法进行专利的标引词识别,通过测度专利标引词在企业文本的映射程度来实现专利与企业相关性的自动量化。本研究对环保企业专利与企业相关性做整体评价判断,在任意选取的4325份环保领域专利中,有4065个一致性结果,260个不一致结果,其准确率达到94%,验证了本文方法的可行性。
With the number of patent applications in China growing each year,some of these applications are becoming increasingly chaotic,as evident from the actions of certain enterprises in declaring or purchasing patents indiscriminately in order to win policy awards under the state’s support and encouragement policies for patent applications.This can result in patents that have no positive impact on corporate development or profit improvement,and that are not related to the company’s main products.In this context,patent and enterprise relevance measurement research has become a valuable research topic.In considering the subjective evaluation methods of experts in the field of patent and enterprise relevance judgment,this study proposes a more valuable,innovative application method and provides an alternative to the traditional subjective evaluation methods.Information theory and deep learning methods are used to identify the index of patents.Additionally,the degree to which patent index words are mapped in corporate texts is measured in order to achieve an automatic quantification of patent and enterprise relevance.This study conducts an overall evaluation of the relevance of patents and enterprises in the environmental protection industry.Among the 4325 patents in the environmental protection field,there are 4065 consistent results and 260 inconsistent results,or an accuracy rate of 94%.The feasibility of the method is thus verified.
作者
高影繁
王峥
胡小荣
姚长青
梁娜
Gao Yingfan;Wang Zheng;Hu Xiaorong;Yao Changqing;Liang Na(Institute of Scientific and Technical Information of China,Beijing 100038;National Science Library,Chinese Academy of Sciences,Beijing 100190)
出处
《情报学报》
CSSCI
CSCD
北大核心
2019年第12期1268-1274,共7页
Journal of the China Society for Scientific and Technical Information
基金
中国科学技术信息研究所重点工作项目“上市公司年报数据库建设及服务系统研发”（ZD2019-09）
关键词
专利与企业相关性
短语抽取
词向量过滤
标引词权重
patent and enterprise relevance
phrase extraction
filtering by word vector
weights for indexing words
