双层嵌套聚类下的专利群落确定及技术布局策略研究

Research on Patent Community Determination and Technology Layout Strategy Under Double-Layer Nested Clustering

[研究目的]专利群落符合数字化下超级专利网络复杂创新要求,专利群落的科学划分及基于网络特征的技术布局策略研究,可为行业技术发展提供更具针对性、客观性与系统性的决策支持。[研究方法]该文提出双层嵌套聚类确定专利群落的方法。首先,基于IRM并行模型,构建专利种群网络;其次,立足于网络结构、关系以及表型三维度,设计专利种群网络特征测度指标体系,并利用DSM实现网络特征指标聚类,确定专利群落特征集及具体构成;最后,以碳纤维领域为实证对象,识别专利群落,并综合TF-IDF与LDA模型确定各群落技术领域主题,提出技术布局策略。[研究结论]双层嵌套聚类方法能够科学划分专利群落,融合突破、迭代成长与边缘搜索三大群落应分别通过纵横整合、代际储备、利基挖掘等方式实现有效技术布局。

[Research purpose]The patent community meets the complex innovation requirements of the super patent network under digitization.The scientific division of the patent community and the research on the technology layout strategy based on network characteristics can provide more targeted,objective and systematic decision support for the development of industry technology.[Research method]A method for patent community determination was designed in this paper by double-layer nested clustering.Firstly,based on IRM parallel clustering model,patent population network was constructed.Secondly,based on the three dimensions of network structure,network relationship and network phenotype,a patent population network characteristic measure index system was designed.DSM was used to realize the clustering of network characteristic indexes,and to determine the characteristic set and specific composition of the patent community.Finally,taking the carbon fiber field as the research object,the patent communities were identified,and the TF-IDF and LDA models were integrated to determine the technical field themes of each patent community,and the technology layout strategy was proposed.[Research conclusion]The double-layer nested clustering method can scientifically divide the patent community,and the three major communities of fusion breakthrough,iterative growth and edge search should achieve effective technology layout through vertical and horizontal integration,inter-generational reserves,and niche mining.