摘要
为了促进同一学术领域的科研合作团队的组建,提高科研效率,本文基于网络表示学习对多个领域科研合作推荐模型进行研究。将基于节点位置的网络表示学习模型与融合网络结构的网络表示学习模型进行集成,得到新的顶点表示,对两个顶点的表示进行选择二元运算得到边的表示。模型将网络表示学习与机器学习相结合,将节点对的表示作为特征训练逻辑分类器,分类器得到的标签即为链接预测结果。通过对金融和物理领域的论文合作数据进行分析,构建科研合作网络。实验证明,提出的集成模型在AUC值上的表现比单一模型更好,效果最高提升了2.3%;在训练集规模较小的情况下,AUC值仍能达到60%。实验结果表明,该科研合作推荐模型具有可行性,对同一学术领域的科研合作团队的组建能够起到有效辅助作用。
This paper researched a scientific collaboration recommendation model in the financial field based on network embedding to promote the formation of a research team in the same research field and improve the efficiency of research.The model integrates two types of network embedding models;one of these is based on the location of vertices, while the other is integrated with network structure. A binary operator for the representation of two vertices was employed to gener‐ate a representation of edge. Combining network embedding and machine learning, the model trained a logic regression classifier with the representation of edges as features, and the labels acquired from the classifier were the results of link pre‐diction. By analyzing papers in the financial and physical research fields, several scientific collaboration networks were constructed. The experiments confirm that the proposed integrated model has achieved better performance than single mod‐els on the value of AUC, with the efficiency improved by up to 2%;even on a small training set, the value of AUC still reached 60%. The proposed model proved to be feasible in scientific collaboration recommendation, which will effectively promote the formation of a research team in the same field.
作者
余传明
林奥琛
钟韵辞
安璐
Yu Chuanming;Lin Aochen;Zhong Yunci;An Lu(School of Information and Safety Engineering, Zhongnan University of Economics and Law, Wuhan 430073;School of Information, Wuhan University, Wuhan 430072)
出处
《情报学报》
CSSCI
CSCD
北大核心
2019年第5期500-511,共12页
Journal of the China Society for Scientific and Technical Information
基金
国家自然科学基金面上项目"大数据环境下基于领域知识获取与对齐的观点检索研究"(71373286)
关键词
科研合作推荐
链接预测
网络表示学习
机器学习
深度学习
scientific collaboration recommendation
link prediction
network embedding
machine learning
deep learning
