基于多维特征测度的人工智能领域研究前沿分析  

Research Fronts Analysis of AI Based on Multidimensional Feature Measure

作　　者：邓启平[1] 陈卫静[1] 张玲玲[1] 张宇娥[1] Deng Qiping;Chen Weijing;Zhang Lingling;Zhang Yu'e(Library,University of Electronic Science and Technology of China,Chengdu 611731)

机构地区：[1]电子科技大学图书馆

出　　处：《情报杂志》2020年第3期56-62,共7页Journal of Intelligence

基　　金：电子科技大学经常性专项项目“双一流视角下基于ESI的中国高校学科评估研究(三期)”(编号:Y03019036201003)研究成果之一

摘　　要：[目的/意义]人工智能技术为全球经济发展注入了新动力,是国家科技战略发展的重点领域。旨在探析人工智能领域的研究前沿及其未来发展态势,并为其他研究前沿分析工作提供参考。[方法/过程]以中国计算机协会推荐的人工智能领域A类、B类期刊和会议为数据源,利用复杂网络社区探测算法从领域文献的耦合网络中探测研究主题,在此基础上,设计文献计量指标对研究主题的新颖性、增长性和影响力等多维特征测度,并结合指标阈值遴选研究前沿。[结果/结论]从人工智能领域文献集中识别出20个研究前沿并分析其发展态势,将识别结果与已有研究成果进行对比分析,验证了基于多维特征测度的方法能有效识别人工智能领域的研究前沿,对其他领域的分析具有一定的借鉴意义。[Purpose/Significance]AI technology has provided new impetus for the development of global economy, and it has become the key domain of the national science and technology development strategy. This research aims to analyze the research fronts of AI and their development trend, and provides reference for relevant researches.[Method/Process]Journals and conferences of level A and B in the field of AI recommended by China Computer Federation(CCF) were selected as the data source, and the community detection algorithm was used to recognize research topics from bibliometric coupling network of literature collection. Then, multidimensional features such as the novelty, growth and impact of research topic were measured using bibliometric indicators, and research fronts were identified by using thresholds of above indicators.[Result/Conclusion]Twenty research fronts were identified from literature collection in AI, and the development trend of research fronts was analyzed further. By comparing with previous analyses, the method based on multidimensional feature measure was proved to be effective in identifying research fronts of AI, and it also has certain reference significance for the analysis in other fields.

关 键 词：人工智能 文献耦合 多维特征测度 文献计量指标 指标阈值 

分 类 号：G350[文化科学—情报学]