基于专利主题演化的颠覆性技术识别研究——以类脑智能领域为例  

The Identification of Disruptive Technology Based on Patent Theme Evolution:Taking the Field of Brain-Inspired Intelligence as an Example

作　　者：邢晓昭[1] 任亮 雷孝平[1] 刘雨农 XING Xiaozhao;REN Liang;LEI Xiaoping;LIU Yunong(Institute of Scientific and Technological Information of China,Beijing 100083,China)

机构地区：[1]中国科学技术信息研究所,北京100083

出　　处：《情报科学》2023年第3期81-88,共8页Information Science

基　　金：国家社科基金青年项目“基于多源知识网络的颠覆性技术分类识别方法研究”(21CTQ039)。

摘　　要：【目的/意义】颠覆性技术具有隐蔽性和突变性,而技术主题在演化过程中会释放颠覆性趋势产生的早期信号,本文提出一种基于专利主题演化的颠覆性技术识别方法。【方法/过程】首先,基于S曲线法判断技术生命周期;其次,综合运用文本挖掘和社会网络分析方法识别技术主题的演化事件;最后,定义新主题出现、原主题趋热和多主题融合三类突变情况,用于识别具有颠覆性潜力的技术主题。【结果/结论】通过类脑智能领域专利实证,发现六个新出现主题、四个趋热型主题和六个融合型主题,通过专家判读,证明了所提方法的有效性,同时筛选出五项主题作为最终预测结果。【创新/局限】本文基于动态社区发现算法对技术主题的演化特征进行分析,进而基于主题突变性筛选颠覆性技术,但数据来源较为单一,将在未来研究中加以扩充。【Purpose/significance】Disruptive technologies are concealed and mutated,and technological topics will release early signals of disruptive trends during the evolution process.This paper proposes a method to identify disruptive technology based on the evolution of patent themes.【Method/process】Firstly,the technology life cycle is judged based on the S-curve method;Secondly,methods on text mining and social network analysis are adopted to identify the evolutional events of technological topics;finally,three types of mutation situations are defined aim at selecting technological topics with high disruptive potential.【Result/conclusion】Empirical study on patents of brain-inspired intelligence has discovered six emerging topics,four hot ting topics and six merging topics.Through expert interpretation,the effectiveness of the proposed method is proved,and five topics are screened out as the final prediction results.【Innovation/limitation】This paper analyzes the evolutionary characteristics of technical topics based on the dynamic community discovery algorithm and then screens disruptive technologies based on topic mutation.However,the data source is relatively single and will be expanded in future research.

关 键 词：颠覆性技术识别 主题演化 动态社区发现 社会网络分析 专利文本挖掘 

分 类 号：G250.2[文化科学—图书馆学]