摘要
为有效解决基于学科目录进行专家遴选方法的不足,提出一种面向多领域跨学科的专家遴选算法CD-Selection。将专家的研究方向关键词作为专家特征,使用Word2vec算法对论文与专家的研究方向关键词分别进行词语向量化;基于欧氏距离计算论文与专家研究方向关键词的词向量相似度;基于论文研究方向关键词的TF-IDF统计值,结合论文与专家的相似度计算专家匹配度,实现多领域跨学科的专家遴选。实验采用Aminer系统抽取的1043个专家的数据集,其结果表明,CD-Selection算法专家遴选匹配率达到90%以上。
To effectively solve the shortages of the traditional method of selecting experts based on the subject catalog,an expert selection algorithm for crossed disciplines in multiple fields was proposed.The experts’interdisciplinary research direction keywords were taken as expert features and Word2vec algorithm was used to vectorize the keywords of the paper and experts’research direction separately.The similarity between the keyword vectors of the paper and experts’research direction was calculated based on the Euclidean distance.The expert matching degree was calculated based on the combination of the TF-IDF statistical value of paper’s keywords and the similarity between paper and experts,so as to achieve expert selection in multiple fields.The data set of 1043 experts extracted through the Aminer system was used in the experiment.The results of the CD-Selection algorithm show that the match accuracy of expert selection is over 90%.
作者
陈敏璇
戴欢
高玉建
付保川
王金鹏
CHEN Min-xuan;DAI Huan;GAO Yu-jian;FU Bao-chuan;WAN Jin-peng(School of Electronic and Information Engineering,Suzhou University of Science and Technology,Suzhou 215000,China;Information Office,China Academic Degrees and Graduate Education Development Center,Beijing 100083,China)
出处
《计算机工程与设计》
北大核心
2022年第6期1671-1677,共7页
Computer Engineering and Design
基金
国家自然科学基金项目(61702354、61876121)
苏州科技大学科研基金项目(XKZ2017004)
研究生科研创新计划基金项目(SKSJ18_012、SJCX19_0963)
苏州科技大学教改基金项目(SKJG18_05)。
关键词
专家遴选
跨学科
研究方向
词语向量化
Aminer系统
expert selection
crossed discipline
research direction
word vectorization
Aminer system
作者简介
陈敏璇(1997),女,江西南昌人,硕士研究生,研究方向为数据挖掘、机器学习;通讯作者:戴欢(1983),男,江苏镇江人,博士,副教授,CCF专业会员,研究方向为数据挖掘、机器学习,E-mail:daihuanjob@163.com;高玉建(1983),男,河北石家庄人,博士,研究方向为大数据;付保川(1964),男,河南巩义人,博士,教授,研究方向为大数据、复杂系统分析;王金鹏(1998),男,江苏盐城人,硕士研究生,研究方向为数据挖掘、机器学习。
