基于动态主题网络的新兴技术主题识别——以氢燃料电池领域为例  

Emerging Technology Topics Identification Based on Dynamic Topic Network: A Case Study of Hydrogen Fuel Cell

作　　者：慎金花[1] 王薇 张更平[1] 陈红艺 Shen Jinhua;Wang Wei;Zhang Gengping;Chen Hongyi(Tongji University Library,Shanghai 200092;School of Economics and Management,Tongji University,Shanghai 200092)

机构地区：[1]同济大学图书馆,上海200092 [2]同济大学经济与管理学院,上海200092

出　　处：《情报杂志》2024年第9期92-100,共9页Journal of Intelligence

摘　　要：[研究目的]新兴技术代表着未来的技术发展方向,是抢占科技前沿制高点必须把握的关键技术,准确识别新兴技术对国家发展具有重要战略意义。[研究方法]综合专利文本信息和分类信息,构建SBERT-LDA-IPC模型,识别各时间段的技术主题;根据主题相似度矩阵绘制动态主题网络,识别具有创新性和连续性的主题为候选主题,评估候选主题的新颖性和影响力,确定新兴技术主题;以氢燃料电池领域为例进行实证检验。[研究结论]研究表明,SBERT-LDA-IPC模型提高了主题聚类的连贯性和准确性,结合国家在氢燃料电池产业发布的系列政策作为验证依据,识别出的三个新兴技术主题,与国家政策制定和产业发展方向一致。[Research purpose]Emerging technology represents the future direction of technological development,and is the key technology that must be grasped to seize the commanding heights of the frontier of science and technology.The accurate identification of emerging technology is of great significance.[Research method]First of all,the SBERT-LDA-IPC model is constructed using patent text and classification information to identify technology topics in each period.After that,the research draws the dynamic topic network according to the topic similarity matrices and identifies the topics with innovation and continuity as candidate topics.Then the research evaluates the novelty and influence of candidate topics to identify emerging technology topics.Finally,through the empirical research of hydrogen fuel cell patents,the viability and efficacy of the emerging technology topics identification approach described in this research are confirmed.[Research conclusion]The result shows that the SBERT-LDA-IPC model improves the consistency and accuracy of topic clustering.Based on the policies released by the country in the hydrogen fuel cell industry,the three emerging technology topics identified are consistent with the national policy formulation and industrial development direction.

关 键 词：动态主题网络 主题演化 主题识别 专利信息 新兴技术 SBERT-LDA-IPC模型 氢燃料电池 

分 类 号：G353.1[文化科学—情报学]