核心专利识别的共引网络分析研究——以“平坦化工艺”为例

Co-citation Network Analysis for Core Patent Recognition: A Case Study of Planarization Process

[目的/意义]探索利用专利共引网络识别核心专利的可行性及其方法,发展专利网络分析理论和方法。[方法/过程]首先,构建专利共引和直接引证网络,比较网络参数,分析引用时滞分布差异;其次,定义节点重要性评价指标CPTR;最后,利用CPTR识别核心专利,比较两种网络识别结果。[结果/结论]案例研究发现,专利共引网络和引证网络存在差异,共引网络能够展现专利关系的多样性特征,具有更显著的整体结构和更好的“知识流”传递能力;CPTR指标降低了PageRank算法的约束条件,识别的“平坦化工艺”技术领域内的核心专利的时效性更强。

[Purpose/significance]This paper explores the feasibility and method of using patent co-citation network to identify core patents and develops patent network analysis theory and method.[Method/process]Firstly,patent co-citation and patent citation networks are constructed,network parameters are compared,and time delay distribution differences are analyzed.Secondly,the node importance evaluation index CPTR of the patent cocitation network is defined and calculated,and the relations hip between its score distribution characteristics and technology value distribution characteristics is compared.Finally,CPTR is used to identify core patents,and the identification results of the two networks are compared.[Result/conclusion]After the empirical analysis of the field of“planarization process”,the results reported in this paper indicate that,there are differences between the patent co-citation network and the citation network.The co-citation network can show the diversity characteristics of patent relations, and has a more significant overall structure and better “knowledge flow” transmission ability;CPTR index reduces the constraint condition of PageRank algorithm, and the core patents identified in the f ield of “planarization process” technology are more timeliness.