Research And Application Of Representation Learning In Social Networks

Nowadays,the development of the Internet has developed rapidly.Social network research is also being studied.Among them,the research on representation learning in social networks occupies an important part.At present,the research on representation learning of social networks is in an in-depth stage.However,the research on representation learning of social networks is more complicated,and there is still space for further research and progress.Network Representation Learning is designed to learn low-dimensional potential representations of nodes in a network that can be used as various feature tasks on the graph,such as classification,clustering,link prediction,node prediction,and visualization.This paper summarizes the representation learning of social networks by summarizing and analyzing the latest developments in this field.This article begins with a brief introduction to the development of representation learning,social networking.Then,the article discusses the shortcomings of the current research methods in the social network,and proposes a social network representation learning research model for the attribute graph,and compares the algorithm with the previous algorithm.Finally,this paper further designs the application platform of social network representation learning,and tests the platform.Finally,this paper summarizes the research direction of future work in this field.In summary,the work done in this paper has the following three points:(1)Analyze and summarize the predecessors’ three aspects of representation learning,social networks and representation learning in social networks,and analyze the existing development status.This provides basic knowledge introduction and theoretical research basis for our research.(2)For the representation learning problem in social networks,we construct a attribute graph social network representation learning model.The video + audio algorithm is used to transform the content + structure model for social network data.Because the amount of social network data is relatively large and has attribute information,when using traditional algorithm calculation,it usually can not achieve satisfactory results.For this reason,this paper proposes improved model calculation.In the improved algorithm,the content + structure attribute graph model is mainly designed.Experimental results and comparative analysis show that the designed algorithm can effectively solve the problem.(3)UML modeling technology is used to analyze the system function requirements.The results show that the system includes system setting management,social network learning management,database management and analysis decision management.UML modeling class diagrams and timing diagrams are used for system function design.The user influence algorithm is written in C++ as the system kernel,and the system is written in C#.