Specific Group Discovery Based On Network Gene Theory

With the rapid development of Internet technology,its proportion in people's production and life is gradually increasing,and the scale of Internet users is also expanding year by year,especially social network applications,which gathers almost most of the Internet users,people's social life has gradually shifted from the real world to the virtual network world,presenting a situation of virtual and real mapping.In social networks,because of some internal factors that are the same or similar,such as interests and positions,a group that is organized together is called a group.Individuals within a group can interact and share information,and then influence each other.The discovery of social network groups is of great significance to real-world work.On the one hand,group discovery can be widely used in commercial applications,especially in retrieval,communication,and recommendation.It can provide strong support;on the other hand,Because today's groups rely on virtual Internet technology,a large number of groups that gather for the purpose of malicious behavior are gathered on the Internet,such as the public opinion navy and rumor groups.These groups gather for the same malicious purpose and have the same behavior pattern.Serious Affecting or even jeopardizing the security of the Internet and the security of the real world,such groups need to be discovered and controlled in a timely and effective manner.Therefore,the discovery of a specific group is a very practical topic.Because the users in the Internet are often manipulated by individuals in the real world,the behavior of Internet users in the network usually has the behavior characteristics of real users.Through mining,these characteristics can uniquely characterize a certain subject in the Internet,drawing on biological The concept of genes in science produces the theory of network genes.Our subject is to discover specific groups on the basis of network gene theory.The core of the composition structure of network genes is the network gene unit,which can determine the characteristics of the main body of the network,and can characterize the essential characteristics of a certain aspect of the network entity.Therefore,the network gene unit is mainly used in the research of group discovery.The involved gene units include topical gene units and positional gene units.Network subjects who hold the same or similar positions on the same topic on the network belong to the same group.When the number of network subjects found reaches a certain level,it means that a group is found.Therefore,this article needs to extract topic gene units from the network tweets published by the subject Gene unit with position.Topics are extracted by methods such as LSA,and the similarity between topic gene units is described based on TF-IDF and cosine vectors,thereby merging topics,and extracting position gene units by sentiment analysis.On the basis of the discovered groups,this paper further identifies key people,by extracting topic influence gene units and personal influence gene units,and synthesizing two indicators to identify key people,that is,network subjects with greater influence.