Research On Dynamic Information Propagation Model Based On Nodes And Information Attributes

The rapid development of social network has changed people's traditional social way,it has become the main platform for people to obtain information and exchange opinions.The internal propagation characteristics and mechanism of network information can be revealed by studying the characteristics of information propagation on online social networks.It is of great significance to maintain national network security and creat market value.Based on the dynamics theory of information propagation,this thesis studied the driving factors of single information propagation and the influence of dual information interactions on propagation.The main research contents are as follows.(1)This thesis explored the influence of different factors on information propagation from two aspects of nodes attributes and information attributes.First,based on SIR model,the ‘known' nodes' state was introuced to distinguish the direct and indirect communicators.Secondly,the influence of the background factors and centrality of nodes on the nodes was measured.And the probability of state transition of users with different immunity under the influence of neighbor nodes with different influence was proposed from the perspective of audience and conformity,which considered the interaction behavior of users in the network.Thirdly,considering that the information in the network has its own attributes,so the information attributes were divided into: the matching degree of information content label and user interest label,information influence and information reputation value.This thesis used the logistic regression model to judge the probability of information propagation under the three attributes,so as to define the information value.Considering the division of user's state and the driving effect of node and information on group state transition,V-SKIR dynamic information propagation model based on nodes attributes and information value was established.Finally,the validity of the model was verified by experiments,and the influence of information value,direct rate,indirect rate and recovery probability on the information propagation were explored on the real network.(2)Dual information propagation laws under different functions were studied,from the perspective of information interaction.Firstly,on the basis of SIR model,the nodes in the network that propagate different information were distinguished.Secondly,according to six kinds of information action relations,the transformation coefficient was defined to determine the transfer probability among all kinds of communicators.Thirdly,the priority propagation probability of the dual information was determined by the node state transition probability,which was defined by the nodes and informations attributes.Based on the information interaction,D2-SIR information propagation model was established by considering the information interaction relationship,the information priority rate and the transform between the nodes' propagation states.Finally,the rule of information communication laws under the interaction relationship was proved through experiments.Meanwhile,the influence of node degree,time difference and network density on the dual information communication were explored.This thesis contains 42 pieces of figures,12 pieces of tables and 87 pieces of references.