摘要
把握学科与研究领域的关键主题对于科研工作者和管理者都具有重要意义。本文提出一种基于"半积累"引文网络的社区结构来识别领域主题及主题演化的方法。并以"合作演化"这一学科领域为例,对所提出的方法进行检验。构建这一领域7个相继时间段下的引文网络并通过社区发现算法识别主题,并根据相邻时间段社区之间的引用强度来识别主题演化路径以及不同主题间的交互影响结构。在"合作演化"这一实例下主题发现中展现的优势表明本方法有望同基于文献耦合网络以及同被引网络分析的主题识别与主题演化分析方法结合以深化学科主题挖掘。
The grasping of key research topics in disciplines and academic domains is of great importance for both scientific researchers and policy-makers. A method is proposed in this paper to identify the key research topics and their evolutionary paths through the detection of communities in the ＂partial-cumulative＂ citation network. The proposed method is examined in the research field of ＂evolution of cooperation＂. By constructing seven successive citation networks in this field, the key topics are identified through community detection and the evolutionary paths of topics are tracked by detecting the strong connections between communities in adjacent time-slices; and the interrela- tionships between the identified topics can accordingly be analyzed. The proposed method has better performance in the accuracy in topic identification in the ＂evolution of cooperation＂ dataset in comparison with the methods based on bibliographic coupling networks and co-citation networks, indicating that the proposed method may complement the existing citation-based methods for topic identification and tracking of topic evolution.
出处
《情报学报》
CSSCI
CSCD
北大核心
2017年第1期100-110,共11页
Journal of the China Society for Scientific and Technical Information
基金
国家自然科学基金项目"支持社会化创新的在线社区知识活动动力学研究"(71371040)
国家自然科学基金项目"时变在线社会网络中舆情传播与演化动力学研究"(71401024)
关键词
主题演化
引文网络
社区发现
合作演化
topic evolution
citation network
community detection
evolution of cooperation
