In view of the cold start and information silo problems faced by the user management of multi-source complex networks,it is difficult for enterprises to meet the dilemma of identifying the same user in different social networks.It is also difficult to achieve rapid and accurate user recognition with the traditional cross social net-work user alignment method that only uses a single dimension feature extraction source.The use of multidimen-sional feature extraction sources from the fine-grained perspective can effectively couple the advantages of vari-ous users'information to improve recognition efficiency and accuracy.The paper first conducted a systematic analysis of relevant literature on cross social network user alignment at home and abroad.Starting from the fine-grained perspective,the main content of the literature was summarized and sorted out.Then,a detailed compara-tive analysis was conducted one by one to clarify the differences,advantages,and constraints between different feature extraction sources used in existing research.Finally,the data were obtained from a private dataset.From the perspectives of multidimensional data integration and multi-source social network data coupling,we provide prospects for the cutting-edge research directions and in-depth exploration in the field of user alignment in the future.
关键词
跨社交网络/复杂网络/数据挖掘/用户对齐
Key words
cross social networks/complex networks/data mining/user alignment
维普
万方数据