Group Detection System Based On Social Network Specific User

The rapid development of social networks has attracted a large number of users,which make Internet have become one of the main aggregation of public opinion.In recent years,almost all the major events in all fields all over the world have been discussed on the Internet.Therefore,analyzing the public opinion on social networks has become a hot research topic.Some key characters,which are called "opinion leaders",are very influential in forming and spread of public opinion in micro blog.Opinion leaders usually have strong influences.So they are leading the public opinions about hotspots and focus in their respective fields.When analyzing the public opinion on Internet,it would effectively save the cost if we find out the opinion leaders on the social network and acquire,store and analyze data from them.In order to discover the social network opinion leaders,we developed the user detection system based on social network specific users.This system can detect other opinion leaders in social network with some opinion leaders as its import.And it is also able to acquire user info and message aiming at the detected opinion leaders.In order to decide the opinion leaders list,this system has realized the influential analyzing algorithm and community detection algorithm.It also improves the community detection algorithm according to the system's specific conditions,which improves the precision of dividing incomplete user-relation diagram.As to the data acquisition part,we build a distributed social network data crawler as data acquisition module in the system.The crawler acquires data by two ways:calling the social networks API and simulating browser.It can acquire user information,users relationships,history messages and real time messages.In order to improve the acquisition efficiency,data acquisition module uses distributed technology to make multiple hosts cooperate acquiring data.And each host can switch IP and account timing to break through social network's constraints on the acquisition rate.Finally,this system also show,modify and share the acquired data.The system implementation result proves that all functions have achieved and the system is running efficiently.The user detection system has been formally applied to network public opinion analysis,running in good condition.It can finish user detection and data acquisition jobs stably and efficiently.