Researh Of E-mail Network Security Trend Analysis Based On Information Dynamic Model

Nowadays,e-mails play an important role in people's life and work.The information and relationships of e-mails reflect the behaviors and connection of people in real society,and e-mail security behaviors can reflect the security behaviors of people.Besides,the features and trend of the e-mail network can also reflect the characteristics of people as well as the features of relationships between them in the real world.Therefore,by analyzing the security behaviors of e-mails,the features of behaviors and trend of people in the corresponding group can be indirectly obtained.In order to obtain characteristics and relationships of characters in real society through e-mail networks,this paper does research of analyzing the security trends of e-mail network using information dynamic models.In order to improve validity and accuracy of analysis of e-mail networks'security trend,this paper combines k-shell decomposition with LeaderRank algorithms to make up information dynamic model.The k-shell decomposition information dynamic model aims to analyze and simplify the structure of complex networks.Delete nodes whose dynamic value is smaller than others gradually to reduce the influence of disturbed nodes to network analysis and decomposing the complicated network.But on account of the measure indicators are not enough and the sequence of being deleted is ignored,some relative important nodes are removed.Thus,this paper proposes an improved k-shell decomposition information dynamic model by combining k-core threshold and Ksd threshold to divide network layers to enhance validity of network analysis afterwards.The LeaderRank information dynamic model evaluates the information dynamic value of nodes in complicated networks.It adjusts the dynamic value in original networks by means of adding a background node,which helps to accurately analyze the behaviors.But the adjusting parameters of added background node defined in LeaderRank information dynamic model need to be improved.Therefore,this paper proposes KLA information dynamic model to figure out and analyze email networks.It also proves that KLA information dynamic model proposed in this paper has better accuracy and effectiveness in complicated networks'behaviors and trend analysis through doing experiments.Based on the above research,this paper designs the e-mail security behavior trend analysis system by using the improved information dynamic model proposed in this paper.It can visually display e-mail networks,count and analyze the security behaviors of e-mails to find its security trend.This system can show the characteristics of people in reality and the relationships between them.It can figure out whether the status of characters is normal,which can improve the security of the entity corresponding to the e-mail networks.The results of experiments prove that the information dynamic model proposed in this paper does better than others.