摘要
[目的/意义]精准识别潜在专利技术组合,将相互关联的技术组成保护范围更大的专利网,对于打破单项专利的局限性、构建更为缜密的技术壁垒具有重要意义。[方法/过程]首先在对关键词与核心IPC进行语义抽取的基础上,筛选出核心专利集合,然后计算基于专利相似性与互补性的专利组合强度,最后利用MCL聚类算法直观、精准地识别潜在专利组合,并以艾滋病疫苗领域专利对方法进行了验证。[结果/结论]该方法以核心专利集合为数据源有效降低了组合识别中的噪音,基于多维度的专利组合强度计算克服了以往组合识别指标的片面性,利用MCL聚类算法无需人为规定簇群数量,保证识别质量。
[Purpose/significance] Accurately identifying potential patent technology portfolios and combining related technologies into a patent network with greater protection scope is of great significance for breaking the limitations of individual patents and building more sophisticated technical barriers.[Method/process] Firstly,based on the semantic extraction of keywords and core IPC,the core patent collections are selected,and then the patent portfolio strength based on patent similarity and complementarity is calculated.Finally,the MCL clustering algorithm is used to identify potential patent portfolios intuitively and accurately.The method was validated with patents in the field of AIDS vaccines.[Result/conclusion] This method uses the core patent collection as the data source to effectively reduce the noise in combination recognition.The multi-dimensional patent combination strength calculation overcomes the one-sidedness of the previous recognition index.The MCL clustering algorithm does not need to artificially specify the number of clusters,which can ensure recognition quality.
出处
《情报理论与实践》
CSSCI
北大核心
2021年第2期178-184,共7页
Information Studies:Theory & Application
基金
国家社会科学基金项目“高校图书馆深度嵌入专利运营研究”的成果之一,项目编号:16BTQ029。
关键词
专利技术组合
识别方法
核心专利集
MCL算法
专利筛选
potential patent portfolio
identification method
core patent set
MCL algorithm
patent screening
