基于神经网络的专利技术机会分析——以生物纳米医药为例  

The technology opportunity analysis of patent based on neural network--taking biological nanomedicine as an example

作　　者：单治易 关陟昊 安新颖 SHAN Zhi-yi;CUAN Zhi-hao;AN Xin-ying(Institute of Medical Information,Peking Union Medical College/Chinese Academy of Medical Sciences,Beijing 100020,China;National Science Library,Chinese Academy of Sciences,Beijing 100190,China;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190,China;Agricultural Information Institution,Chinese Academy of Agriculture Sciences,Bejing 100081,China)

机构地区：[1]北京协和医学院/中国医学科学院医学信息研究所,北京100020 [2]中国科学院文献情报中心,北京100190 [3]中国科学院大学经济与管理学院信息资源管理系,北京100190 [4]中国农业科学院农业信息研究所,北京100081

出　　处：《中国新药杂志》2023年第9期865-871,共7页Chinese Journal of New Drugs

基　　金：中国医学科学院医学与健康科技创新工程项目“生物医学文献信息保障与集成服务平台”(2021-I2M-1-033);中国医学科学院中央级公益性科研院所基本科研业务费“卓越导向的医学科技成果评价与激励机制研究”(2022-ZHL630-01)。

摘　　要：目的:本研究以专利数据为基础提出一种基于神经网络的技术机会分析方法。方法:首先,基于深度学习算法学习数据的内容特征,形成抽取关键词模型,基于内容计算关键词间内容相似度特征。其次,基于关键词构建专利技术网络。运用链路预测算法,计算网络的拓扑结构特征,包括:基于局部信息的相似性特征、基于随机游走的相似性特征和基于路径的相似性特征。最后,将3类网络拓扑特征和内容特征输入到反向传播神经网络模型中,预测潜在的技术机会。结果:选择生物纳米医药领域作为实证领域,对生物纳米医药领域专利技术网络中概率得分最大的前10项技术机会进行识别。结论:通过文献回溯法对该结果进行分析评价,通过最新的文献发现学者对这些技术机会的研究成果,但尚未成功申请专利。Objective:This study proposes a neural network-based technology opportunity analysis of patent data.Methods:First,based on learning of the content characteristics of the data by deep learning algorithm,the extraction keyword model was formed.Content similarity between keywords was calculated based on the contents.Second,patent technology network was built based on the keywords.The link prediction algorithm was used to calculate the topology features of the network,including similarity features based on local information,random walk and path.Finally,the three types of network topology features and content features were input to the BP neural network model to predict potential technology opportunities.Results:The field of biological nanomedicine was selected as the empirical field.The top 10 technological opportunities with the highest probability score in the patented technology network were identified in the field of biomedicine.Conclusion:The results were analyzed and evaluated through the literature retrospective method,and the latest literature found the research results on these technical opportunities,but they have not successfully applied for patents.

关 键 词：技术机会 链路预测 复杂网络 反向传播神经网络 

分 类 号：R95[医药卫生—药学]