Division And Visualization Of Overlapping Internet Community

As one of the most representative information dissenmination platforms,micro-blog has become an important research object in many disciplines,such as computer science and sociology.Using social network analysis method to study users' behaviors and detect communities is helpful for mastering user aggregation,so as to standardize network management,guide and monitor cyber group events,and is of great significance for maintaining network security.Taking micro-blog user behaviors and network as the research object,aiming at the main problems existing problems in this research,the thesis focuses on micro-blog users' behaviors,micro-blog overlapping community detection,and micro-blog community visualization layout.First,since existing researches are lack of comprehensive users' behaviors quantification,a new users' behavioral relationship tightness algorithm is proposed.The algorithm considers the difference between users' behaviors,calculates the proportion of forwarding and mentioning number to the total number of corresponding behaviors between each other.Then,the closeness of user's behaviors is calculated by combining attention,forwarding,and mentioning behaviors.The experiment chooses authentic micro-blog data.Experiment results show that the improved algorithm can effectively reflect the closeness of users.Combined the users' behavioral relationship tightness with the recent activity,the user importance contribution is calculated.Then,an improved micro-blog user importance ranking algorithm is proposed.Experiment results show that the algorithm can effectively identify highly active users in recent days.Secondly,since the existing micro-blog community detection researches apply the traditional clustering algorithm and the result of the detection cannot reflect user behaviors,a new micro-blog network overlapping community detection algorithm that combines user behaviors is proposed.The users' behaviors relationship tightness is integrated into the greedy group expansion overlapping community detection algorithm,and the community fitness function is changed for the weighted network.Experiment results show that the improved algorithm is more adaptable to the characteristics of micro-blog,and the community structure can better reflect the closeness of users' behaviors.Thirdly,among the existing researches of community visualization,the layout algorithm can poorly display the community structure.Aiming at this problem a new layout algorithm that combines the characteristics of micro-blog network is proposed.The algorithm optimizes the FR algorithm by adding the tightness of users' behaviors.In addition,the community gravity force is added to achieve the internal clustering.Compared with the previous micro-blog network visualization methods,experiment results show that community structure is clear and the algorithm is more efficient by using the proposed layout algorithm.