首页 | 本学科首页   官方微博 | 高级检索  
     检索      

专利前沿技术主题识别及趋势预测方法——以人工智能领域为例
引用本文:宋凯,朱彦君.专利前沿技术主题识别及趋势预测方法——以人工智能领域为例[J].情报杂志,2021(1):33-38.
作者姓名:宋凯  朱彦君
作者单位:武汉大学信息管理学院;山东建筑大学图书馆
摘    要:目的/意义]产业变革快速演进,技术创新成为驱动社会经济发展、提高国家和企业科技竞争力的关键所在,如何对前沿技术进行识别和预测,成为国家科技政策研究和企业技术创新活动关注的热点。方法/过程]以人工智能作为重点研究领域,首先以LDA模型进行技术主题抽取,并结合K-means算法进行专利文本聚类;在此基础上,以Z分数表示技术主题创新度,以Sen's斜率估计技术主题授权趋势,两个指标结合形成技术主题前沿度并将二者映射到二维空间,识别前沿技术主题以及划分技术主题类型;再,计算前沿技术主题的新颖度和关注度,二者融合形成技术主题趋势度指标;最后,采用三次指数平滑法对前沿技术主题的发展趋势进行预测。结果/结论]人工智能领域的前沿技术主题有“智能家居”“电动汽车”和“自动化控制系统”,其中“智能家居”在未来3年的发展呈下降态势,而“电动汽车”和“自动化控制系统”的发展呈明显上升趋势。

关 键 词:专利前沿技术  主题模型  主题识别  人工智能  趋势预测

Patent Frontier Technology Topic Identification and Trend Prediction:A Case Analysis of Artificial Intelligence
Song Kai,Zhu Yanjun.Patent Frontier Technology Topic Identification and Trend Prediction:A Case Analysis of Artificial Intelligence[J].Journal of Information,2021(1):33-38.
Authors:Song Kai  Zhu Yanjun
Institution:(School of Information Management,Wuhan University,Wuhan 430072;Library,Shandong Jianzhu University,Jinan 250101)
Abstract:Purpose/Significance]With the rapid evolution of industrial transformation,technological innovation has become the key to driving social and economic development and improving the scientific and technological competitiveness of the country and enterprises.How to identify and predict frontier technologies has become a hot topic in national technology policy research and enterprise technology innovation activities.Method/Process]This paper takes artificial intelligence as the key research area.Firstly,technical topic extraction is carried out by the LDA model,and patent text clustering is carried out by combining with the K-means algorithm.On this basis,the Z score is used to represent the innovation degree of the technology topic,and Sen's slope is used to represent the authorization trend of the technology topic.The two indexes are fused to form the frontier index of the technology topic,and the two indexes are mapped to the two-dimensional space to identify the frontier technology topic and classify the technology topic types.Furthermore,the novelty and attention index of the frontier technology topic was calculated.The combination of the two index forms the trend degree of the technical topic.Finally,the third exponential smoothing method is used to predict the development trend of frontier technology.Result/Conclusion]The frontier topics in artificial intelligence include"smart homes""electric cars"and"automated control systems".Among them,the development trend of"smart home"in the next three years shows a downward trend,while the development trend of"electric vehicle"and"automatic control system"shows an obvious upward trend.
Keywords:patent frontier technology  topic model  topic identification  AI  trend prediction
本文献已被 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号