基于共词网络的专家技能挖掘方法  

A Expert Skill Mining Method Based on Co-word Network

作　　者：邵明阳 单菁[1,2] 王佳英 SHAO Mingyang;SHAN Jing;WANG Jiaying(School of Information and Control Engineering,Shenyang Jianzhu University,Shenyang 110000;School of Software,Shenyang University of Technology,Shenyang 110000)

机构地区：[1]沈阳建筑大学信息与控制工程学院,沈阳110000 [2]沈阳工业大学软件学院,沈阳110000

出　　处：《计算机与数字工程》2024年第7期2129-2134,共6页Computer & Digital Engineering

基　　金：辽宁省教育厅基金项目“面向生物大数据的局部近似匹配技术研究”(编号:lnqn202015)资助。

摘　　要：专家技能指专家熟悉的研究领域,为了更方便快捷地挖掘专家技能,提出了一种基于共词网络的技能挖掘方法。方法从专家发表的文献题目入手,将题目分解成短语,利用短语之间的共现关系构建共词网络,然后将网络划分出不同的技能主题社区,最后根据专家论文题目中所涉及到的社区以及专家在文献中所处的署名位序信息加权得出专家对技能的熟悉程度。实验数据采用DBLP开源数据集,选取发文量大于40篇的300位专家共36981条题目进行验证分析。实验结果表明,基于共词网络的专家技能挖掘方法在准确率和召回率上的表现较好,最高达到了61.3%和69.2%。基于共词网络的专家技能挖掘方法不仅能快速有效处理大量数据,而且可以较好地从专家发表的文献题目中挖掘出专家技能。Expert skills refer to the research fields that experts are familiar with.In order to mine expert skills more conve⁃niently and quickly,a skill mining method based on Co-word network is proposed.Starting with the literature topics published by experts,the topics are decomposed into phrases,and the co-occurrence relationship between phrases is used to construct a Co-word network.Then the network is divided into different skill subject communities.Finally,the familiarity of experts with skills is weighted according to the communities involved in the topics of expert papers and the signature order information of experts in the literature.The experimental data adopts DBLP open source data set,and 300 experts with a total of 36981 topics with more than 40 articles are selected for verification and analysis.The experimental results show that the expert skill mining method based on Co-word network performs well in accuracy and recall,up to 61.3%and 69.2%.The expert skill mining method based on Co-word network can not only deal with a large amount of data quickly and effectively,but also mine expert skills from the literature topics published by experts.

关 键 词：共词网络 社区划分 词语共现 技能挖掘 主题社区 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]