Algorithm Design And System Implementation Of Monitoring Node Selection In Social Network

New messages are published at all times in the social network,and the message is transmitted to form a message propagation event.In the research of social networks,how to select monitoring nodes and effectively monitor the spread of messages is an important issue in maintaining the network order.The monitoring nodes of social networks refer to the key nodes which can reflect the degree of message propagation on the network to a great extent compared to other nodes of social networks.Nowadays,the social network is so large that it is impossible to directly obtain and deal with the specific network structure.Selecting monitoring nodes in this context becomes a more difficult problem.This thesis consists of two parts to study the algorithm design for selecting monitoring nodes and the system implementation.The first part is designing the algorithm for selecting monitoring nodes.This thesis makes full use of methods such as data mining,probability graph model and modular functions to generate frequent sets,build Markov networks,and select monitoring nodes by calculating the optimal value of module functions.The second part is implementing a system for selecting monitoring nodes.By analyzing the process of the system and designing the functional modules of the system,we implement a Web application for selecting monitoring nodes with a large degree of message propagation in the social networks.According to the process of designing the algorithm for selecting monitoring nodes in social networks and implementing the system,the main contributions of this thesis are summarized as follows.We propose a method for selecting monitoring nodes in social networks.The specific contents of the method are: building a Markov Network to ensure all nodes to be selected which are in a connected graph on the basis of the network node relationship;reserving the frequent item sets where the nodes to be selected to process the frequent item sets;introducing a sub-modular function,and selecting the monitoring nodes by calculating the optimal value of sub-modular functions.We do experiments to test the algorithm execution effectiveness of selecting monitoring nodes in social networks.The experiments select micro-blog data sets andTwitter data sets to test the algorithm execution efficiency of selecting monitoring nodes.Then we analyze the main factors which affecting the algorithm execution time.We design a system of selecting monitoring nodes in social networks.First,we carry out the overall design of the system.According to the system processing flow,we divide the system into six functional modules which are uploading data,processing data,mining association rule among network nodes,building Markov networks,and selecting monitoring nodes.Second,according to the division of different module functions,we give out the data flow chart of each module.We implement a system of selecting monitoring nodes in social networks to visualize the results.According to the design requirements of the system including data structures,algorithm implementation,module implementation,etc.,we use programming languages such as JAVA and HTML,and programming framework such as Spring MVC to achieve the system,and use Tomcat server to release system.