情报科学 ›› 2024, Vol. 42 ›› Issue (11): 120-127.

• 业务研究 • 上一篇    下一篇

综合属性和关系的专利网络社区发现和核心专利识别研究

  

  • 出版日期:2024-11-01 发布日期:2025-04-08

  • Online:2024-11-01 Published:2025-04-08

摘要: 【目的/意义】以信息论为理论基础,提出综合属性和关系的专利网络社区发现和核心专利识别方法,旨在 平衡专利的个体特质性和网络整体性。【方法/过程】在构建专利引用网络的基础上,利用德温特手工代码和关键词 参与Walktrap社区发现算法,实现专利技术领域细分。在技术领域细分条件下,综合属性和关系构建专利识别指标 体系,量化表征专利重要性并以此为依据识别核心专利。【结果/结论】实证表明:该方法兼具专利识别指标体系和 网络分析技术关系的优点;在细分技术领域的条件下,识别结果更加精准。【创新/局限】囿于研究内容等原因,仅从 引用关系角度探讨了综合属性和关系的专利网络分析方法,未来将继续探究其他关系的影响。

Abstract: 【Purpose/significance】Based on information theory, this paper puts forward a method of patent network community discov⁃ ery and core patent identification, which integrates attributes and relationships, aiming at balancing the individual characteristics of patents and the integrity of the network【. Method/process】On the basis of constructing patent citation network, Derwent manual code and keywords were used to participate in Walktrap community discovery algorithm, and the patent technology field was subdivided. Under the condition of technical field subdivision, the patent identification index system is constructed by integrating attributes and re⁃ lationships, which quantifies the importance of patents and identifies core patents based on it.【Result/conclusion】The empirical re⁃ sults show that: This method has the advantages of both patent identification index system and network analysis technology relation⁃ ship; Under the condition of subdividing technical fields, the recognition results are more accurate【. Innovation/limitation】Due to the research content and other reasons, this paper only discusses the patent network analysis method of comprehensive attributes and rela⁃ tionships from the perspective of citation relationship, and will continue to explore the influence of other relationships in the future.