异构信息网络下的特征向量中心性排名研究

Research on Eigenvector Centrality Ranking under Heterogeneous Information Network

摘要: 本文从异构网络角度来进行排名的研究，传统的排名研究是基于同构网络来进行的，然而在现实生活中，实际的网络是异构的。首先，本文根据真实的电影数据集来建立一个异构网络后，根据网络得出相应的网络模式，之后根据电影网络中演员、导演、电影之间的相互关系来建立元路径。其次，结合元路径设定相应的排名规则，结合特征向量中心性度量方法来进行排名研究。最后，给出了实验的结果。本文的创新之处在于，将特征向量中心性方法应用到了一个异构网络中，能更好地反映出真实网络的特征，并与度中心性方法进行对比，展示了本文方法具有良好的可行性。

Abstract: This paper studies the ranking from the perspective of heterogeneous networks. Traditional rank-ing research is based on homogeneous networks. However, in real life, the actual networks are heterogeneous. Firstly, after establishing a heterogeneous network based on the real movie dataset, the paper derives the corresponding network model according to the network, and then establishes the meta-path according to the relationship between actors, directors and movies in the movie network. Secondly, the corresponding ranking rules are set in combination with the meta-path, and the eigenvector centrality measurement method is used to conduct the ranking study. Finally, the results of the experiment are given. The innovation of this paper is that the eigenvector centrality method is applied to a heterogeneous network, which can better reflect the characteristics of the real network and compare it with the degree centrality method. It shows that the method has good feasibility.