邱悦文.离群专利与论文视角下新兴技术主题识别研究[J].情报工程,2021,7(3):081-093 |
离群专利与论文视角下新兴技术主题识别研究 |
Research on Emerging Technology Topic Identification based on Outlier Patent and Paper |
|
DOI:10.3772/j.issn.2095-915X.2021.03.007 |
中文关键词: 主题识别;新兴技术;识别方法;车联网 |
英文关键词: Topic identification; emerging technology; dentification Method; IOV |
基金项目:国家重点研发计划项目“基于人工智能技术创新创业服务技术集成研发与应用示范”(2019YFB1405200)。 |
|
摘要点击次数: 2189 |
全文下载次数: 2209 |
中文摘要: |
[ 目的 / 意义 ] 针对当前新兴技术识别主要围绕热点专利与论文开展,对低热度专利与论文不够关注,然而由于其类别较多、总体规模较大,极易产生长尾效应。因此,以低热度专利与论文为研究对象,设计新兴技术主题识别方法,以期为国家、企业、机构提供情报支持。[ 方法 / 过程 ] 构建专利、论文相似矩阵,利用社会网络分析工具聚类筛选出离群专利分类号与论文关键词;通过 LDA 模型获取技术主题;短语相似度测算,归并技术主题;围绕新兴技术特征构建识别指标体系对技术主题进行综合评价,识别出新兴技术。[ 结果 / 结论 ] 以车联网为实验对象,识别出无人驾驶技术是车联网领域新兴技术,识别结果与业界分析一致,说明方法有效性。 |
英文摘要: |
[Objective/ Significance] The topic identification of current emerging technologies is mainly carried out around hot patents and papers, and there is not enough attention to outlier patents and papers. However, these patents and papers are easy to produce long tail effect. Therefore, this research uses outlier patents and papers to design a new method of identifying emerging technology topics, for providing intelligence support for countries, enterprises, and institutions. [Methods/Process] First, construct similarity matrix, use social network analysis to cluster and filter out the outlier patent classification numbers and paper keywords; then use the LDA model to obtain the topic of outlier patents and papers. Next,through the phrase similarity measurement, the technical topics are merged. Finally,the identification system is constructed from characteristics of emerging technology to evaluate the technical topics and identify emerging technologies. [Results/Conclusions] Taking the Internet of Vehicles as the experimental object, it has successfully identified that unmanned driving technology is an emerging technology in the field. The identification result is consistent with the industry analysis, indicating the effectiveness of the method. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |