首页|融合知识网络嵌入特征的潜在诉讼专利预警模型

融合知识网络嵌入特征的潜在诉讼专利预警模型

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
[研究目的]近年来专利诉讼案件频发,专利诉讼风险管理已成为企业技术研发、布局市场需要面对的一大挑战.建立潜在诉讼专利预警模型,能有效帮助创新主体尽早识别出引发诉讼的风险专利,采取应对措施以减少专利诉讼带来的经济损失.[研究方法]提取专利摘要的SAO结构建立技术领域内的知识网络,将目标专利摘要的SAO结构嵌入到先前专利构成的知识网络中计算出相关特征,与专利授权特征融合作为预测诉讼专利的特征集,利用Autogluon机器学习框架完成对潜在诉讼专利的预测.[研究结论]围绕数字信息传输技术开展实证研究,结果表明模型在融合知识网络特征后的预测性能更佳,精准率达到了 76.7％,且特征向量中心性中值、PageRank中值在预测潜在诉讼专利中发挥了重要作用,丰富了当前诉讼专利预测研究的方法体系.
Research on Potential Litigation Patent Early Warning Model Based on Knowledge Network Embedding Characteristics
[Research purpose]Increasing numbers of patent lawsuits were witnessed in recent years and patent litigation risk management has become a major challenge for enterprises in technology research and development and market layout.The establishment of potential liti-gation patent early warning model can effectively help innovative subjects to identify the risk patents that may cause litigation as soon as possible and take countermeasures to reduce the related economic losses.[Research method]The SAO structure of patent abstracts is ex-tracted to establish a knowledge network in the technical field,the SAO structure of target patent abstracts is embedded into the knowledge network composed of previous patents to calculate the relevant features,and the patent grant features are fused as the feature set for predic-ting litigation patents.The Autogluon machine learning framework is used to predict the potential litigation patents.[Research conclu-sion]Empirical research is carried out around digital information transmission technology.The results show that the prediction performance of the model after integrating the features of knowledge network is better,and the accuracy rate reaches 76.7％.Moreover,the median value of feature vector centrality and median PageRank play an important role in predicting the potential litigation patents,which enriches the current method system of litigation patent prediction research.

dynamic knowledge networkpatent litigationpatent early warningSAO structureAutogluonmachine learningdigital information transmission

方曦、彭康、刘云

展开 >

上海应用技术大学经济与管理学院 上海 201418

中国科学院大学公共政策与管理学院 北京 100046

动态知识网络 专利诉讼 专利预警 SAO结构 Autogluon 机器学习 数字信息传输

国家自然科学基金重点国际(地区)合作研究项目科技部创新方法工作专项

718101070042020IM021000

2024

10.3969/j.issn.1002-1965.2024.10.022

情报杂志
陕西省科学技术信息研究所

情报杂志

CSTPCDCSSCICHSSCD北大核心
影响因子：1.502
ISSN：1002-1965
年,卷(期)：2024.43(10)