A Expert Skill Mining Method Based on Co-word Network
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 36 981 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.
Co-word networkcommunity divisionco-occurrence of wordsskill miningtheme communitys
万方数据
维普
