Research On Time-Sensitive Information Dissemination In Online Social Networks

The rapid development of information technology has made social networks more and more important in people's work and life,infiltrating all aspects of people's lives,and research related to social networks has attracted more and more attention.Social network information communication research is one of the areas of concern in social network research.Due to the rapid development of online social networks,the research of information dissemination also has various challenges.The scale and complexity of the topology,coupled with the dynamic variability of the information dissemination process,etc.,are all we are designing.Effective strategies and communication models need to be considered.This thesis conducts research on information dissemination of online social networks in several dimensions: cost budget,time and community:In order to solve the problem of cost minimization information dissemination,this topic proposes a heuristic strategy called MOWC.The algorithm comprehensively considers the static topology and cost budget of the social network.The algorithm first calculates the ratio of the weight of the node to the cost,and then selects the node with the larger ratio as the seed node in turn until the budget is exhausted.We used algorithms such as Random,MaxDegree,and SingleDiscount to experiment with different data sets and cost budgets.It is proved by the experimental results that our proposed algorithm can achieve better results.In an actual online social network,there is a delay in the propagation of information.There are many factors that cause the delay of time.It is more common that the difference in the online time of each user in the social network causes a delay in time.In the process of studying information dissemination,we considered the user's online pattern and designed the TCIO information dissemination model.We first explain the time-sensitive influence maximization problem and prove that it is an NP-hard problem,and the influence function satisfies monotonicity and submodularity.Then the SimpleGreedy algorithm is proposed for this problem.The main idea of this algorithm is to calculate the influence of the node based on the information dissemination model of our design,and then use the greedy choice to select those with maximum marginal effects until the budget is run out.In order to further reduce the computational complexity and optimize the cost choice,we propose a GMAI algorithm,which approximates the margin influence of the node according to the information propagation characteristics of the online social network graph to obtain the seed node set.This paper uses SimpleGreedy algorithm,GMAI algorithm and other algorithms to compare and verify under different datasets.When the budget,information release time and valid time change are obtained,the proposed algorithm can still achieve better results.Finally,we can see through this experiment that time plays an important role in the process of information dissemination.As to social networks with community structure,while taking into account information delay,different community structures have different time slots,and changes in community structure may cause changes in node influence.The general heuristic algorithm may lead to the overlapping problem of influence,and usually does not consider the influence of time on community structure and influence.This topic proposes TCDD algorithm based on the community structure nature of time-sensitive social network.TCDD divides the initial community graph in the social network by the community division algorithm,selects the important community,allocates the seed number according to the size of the community and forms the candidate seed set by the degree discount algorithm.The social network graph of each time slot is formed according to the last time slot fine-tuning.We can obtain the candidate seed nodes of each time slot and summarize them as the total candidate seed nodes,and finally select the final information seed node set according to the budget and cost.In the different datasets,the TCDD algorithm and other algorithms were used in related experiments and the corresponding results were obtained.When the cost budget,information release time and valid time are different,the TCDD algorithm can get relatively good results.