多源信息视角下的多指标新兴技术主题识别研究*
——以智能网联汽车领域为例

唐 恒 邱悦文

(1.江苏大学知识产权学院 镇江 212013；2. 江苏大学科技信息研究所 镇江 212013)

Emerging Technology Topic Identification Based on Multi-Source Information:Intelligent Connected Vehicle as an Example

Tang Heng1Qiu Yuewen2

(1.School of Intellectual Property Jiangsu University, Zhenjiang 212013;2.Institute of Science and Technology Information, Jiangsu University, Zhenjiang 212013)

Abstract:[Purpose/Significance]In order to improve the comprehensiveness and accuracy of the identification of emerging technology topics, it is necessary to enrich data source information. This research proposes a composite emerging technology topic identification method on the basis of patent, paper and public opinion information. It is expected to provide a reference for the research on the topic identification methods of emerging technologies at the micro level.[Method/Process]Firstly, using the LDA model to extract candidate emerging technology topics. Then, choosing the final emerging technology topics by theme screening indicators. Finally, the emerging technology topic is predicted by using CRITIC method.[Result/Conclusion]In the experiment, intelligent connected vehicle is taken as an example. The experiment successfully identifies one emerging technology: multi-sensor fusion information technology. The experimental results are consistent with the industry analysis, indicating that the multi-indicator emerging technology topic identification method from the multi-source information perspective can effectively target multiple sources.

Keywords:topic identification; emerging technology; LDA model; intelligent connected vehicle; EM；CRITIC

0 引 言

新兴技术 (Emerging Technology ) 是以交叉学科为特征，拥有足够大的发展潜力和优势，且存在对传统技术的改造趋势或者对国家经济结构产生深远影响的技术[1]。当今世界，技术飞速发展，技术发展的新领域与方向一定程度上影响着国家制定科技战略布局。论文记录绝大部分基础(理论)创新研究成果，专利则主要记录技术创新信息，二者皆可反映新理论、新技术的创新性，这与新兴技术最重要的特征(根本创新性)吻合，因此可以作为测度新兴技术的重要数据源。舆情数据(公众观点或意见)则是对技术专家观点(主要体现在科技论文和专利文献中)的有益补充[4]。因此，采用情报学的理论与方法，充分挖掘论文、专利、舆情信息背后的隐含信息，不仅能够为国家制定科技战略布局提供有力的情报支持，对指导企业研发新技术也同样意义非凡。

论文、专利、舆情背后的技术信息各有侧重。论文数据主要表现新兴技术的研究基础[2]，专利数据主要反映新兴技术的创新性[3，5]，而舆情数据主要反映大众对技术发展最直接的认知反馈[6]。对于已有的新兴技术的识别工作，有大量学者对其做过详细的综述[7-9]。表明现有研究对新兴技术主题识别的推进，仍存在以下局限：第一，不同数据类型反映的科技信息侧重点不同，如科技论文侧重基础科学研究产出，专利侧重技术创新。只有全面考虑多源信息，科学计量分析才能得到更客观的分析结果。……