研究者出版记录中的会议序列提取研究  

Conference Sequence Extracting in Researchers' Publication Record

作　　者：王金龙[1] 徐从富[2] 

机构地区：[1]青岛理工大学计算机工程学院,山东青岛266033 [2]浙江大学计算机科学与技术学院,杭州310027

出　　处：《电子科技大学学报》2010年第1期110-113,共4页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金(60802066);国家863计划(2007AA01Z197);山东省优秀中青年科学家科研奖励基金(2008BS01009)

摘　　要：对研究者发文序列的研究不仅可帮助了解领域研究现状,而且对于会议评级、热点预测分析等都有重要的意义。通过对研究者出版序列中的会议进行分析,提取会议序列并进行聚类,有助于发现领域相关性,理解领域发展演化。该文针对会议序列提取中直接考虑前后顺序所存在的不准确,提出了根据时间片相关的会议序列提取方法,以提高会议聚类的准确性。通过真实数据集上的定量和定性实验证明了该方法的有效性,并对结果进行了实例分析。Researchers＇ publication sequence contain some important information which can be utilized to discover the research areas trend, reflect the relative ranking of academic conferences, and indicate the hot research topics. A systematic method is proposed to discover the knowledge hidden in the publication sequences to identify the connection between different research disciplines and to understand the evolution of the research area. More specifically, for the inaccuracy of the consecutive conferences sequences extracted from publication sequences, a new method is presented to extract sequences based on time slice information, so that the clustering accuracy can be improved. The quantitative and qualitative analysis of experiments in real datasets has been provided to illustrate the effectiveness of proposed method.

关 键 词：聚类分析 会议序列 研究者发文序列 时间片 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]