摘要
[目的/意义]“整体维度+局部阶段”的双元演化视角下,挖掘工业机器人领域的技术主题及其演化规律,既可了解技术发展整体历程,又可明确技术组合具体范式,对于洞察工业机器人领域的技术进步落脚点和资金投入关注点具有重要的现实意义。[方法/过程]基于incoPat专利数据库,以2003—2022年工业机器人领域为例,结合Word2vec词向量模型与LDA主题模型对专利文本进行数据挖掘与语料库扩充,继而在整体生态层面识别领域技能主题;随后结合技术生命周期理论,对各阶段的技术主题进行深度挖掘(相似度计算、相关性分析、技能组合规律);对挖掘出的45种技术主题路径进行刻画和可视化表达,明确技能主题发展现状及演化规律。[结果/结论]研究表明,文章所提方法流程能够更精准地识别出整体维度视角下的工业机器人领域的多个主题,且能识别出局部阶段视角下的主题侧重。在技术发展过程中,技术主题演化呈现出“扩散化→丰富化→系统化”的变化规律。通过计算演化强度来刻画主路径,也能清晰地展示工业机器人领域关键技术的波动情况。文章所提出的方法和流程能够为工业机器人领域的技术整合、资金投入、发展创新提供有力的理论和实践支撑。
[Purpose/significance]From the perspective of"overall ecology+local stage",mining the technical theme and its evolution law in the field of industrial robots can not only know the overall process of technological development,but also clarify the specific paradigm of technology combination,which has important practical significance for insight into the technological progress and capital investment focus in the field of industrial robots.[Method/process]Based on incoPat patent database,taking the industrial robot field from 2003 to 2022 as an example,combined with Word2vec word vector model and LDA topic model,data mining and corpus expansion of patent texts were carried out,and then domain skill topics were identified at the overall ecological level.Then,combined with the technology life cycle theory,the technical topics of each stage are deeply explored(similarity calculation,correlation analysis,skill combination law).The 45 kinds of technical topic paths are described and visualized,and the development status and evolution law of skill topics are clarified.[Result/conclusion]The research shows that the proposed method process can more accurately identify multiple topics in the field of industrial robots from the perspective of the overall ecology,and can identify the thematic focus from the perspective of local stages.In the process of technology development,the evolution of technology theme shows a change law of"diffusion→enrichment→systematization".By calculating the evolutionary strength to characterize the main path,it can also clearly show the fluctuation of key technologies in the field of industrial robots.The method and process proposed in this paper can provide strong theoretical and practical support for technology integration,capital investment,development and innovation in the field of industrial robots.
作者
窦路遥
周志刚
李毅
蒋涛
DOU Luyao;ZHOU Zhigang;LI Yi;JIANG Tao(Institute of Information,Shaanxi University of Finance and Economics,Taiyuan 030006;Agricultural Bank of China Anhui Branch,Hefei 230000;Institute of Control Science and Engineering,China University of Petroleum,Qingdao 266000)
出处
《科技情报研究》
2024年第1期102-118,共17页
Scientific Information Research
基金
国家自然科学基金项目“多源数据融合场景下的对抗式隐私洞察靶向保护技术研究”(编号:61902226)
山西省哲学社会科学规划课题“山西省中小企业间数据融合安全共享机制研究”(编号:2022YY097)。
关键词
主题挖掘
动态波动
双元演化
专利分析
工业机器人
topic mining
dynamic fluctuation
binary evolution
patent analysis
industrial robot