Research On Key Security Issues Of Information Spreading In Online Social Networks

With the popularization of mobile devices and the development of communication technologies,social network services have become an important platform for people to communicate,participate in activities,and obtain various services in daily life.However,sensitive information of online users,such as identity,home address,bank account,are always stored on mobile phones.Therefore,some attackers or malicious programs can easily steal users' privacy.In serious cases,it can also pose threat to users' personal safety.On the other hand,with the rapid development of information technology,the way people exchange information is faster and more convenient than before.However,false information and rumors also spread faster,deeper,and wider,which can threat personal life,social order,and even political stability of a country.Therefore,accurately revealing the patterns of online rumors and providing effective counter measures to minimize the harm caused by rumors has become one of the hot research topic.Overall,it is of great importance to conduct in-depth research on privacy and security issues of information dissemination in online social networks.This dissertation focuses on the security requirements of social network,especially regards to information dissemination.This study comprehensively conducts research on several key security issues,such as personalized privacy protection,privacy protection while improving data utility,and rumor spreading models.The study also makes use of differential privacy,game theory,information entropy,and related theories.In order to develop effective protecting technologies with respect to key security problems.The main research contents and contributions of this paper are as follows:1.This dissertation studies the personalized differential privacy protection mechanism and proposes a differential privacy protection model based on the trust degrees among social network users.By mapping trust relationship to privacy protection strength,the mechanism can meet personalized requirement of users,at the same time improve the data utility to some extent.Furthermore,an incomplete information game process established to describe the confrontation of users and attackers.We also provide quantitative analysis and measurement of attack in differential privacy scenarios.The proposed method provides an effective way to further balance privacy protection levels and data utility.2.Based on the above research,the privacy protection mechanism in dynamic scenarios is further studied.Combined with the real application scenario,the mathematical model is established according to the Markov decision process,in which the privacy protection problem in the multi-stage dynamic attack scenario is transformed into a finite-phase zero-sum game process,and the payoff function is defined according to the privacy leakage probability of users and the service quality.Then,we can get the best strategy for users by deriving Nash Equilibrium.In addition,based on the developed enhanced learning algorithm,the algorithm can achieve fast convergence by reducing the cardinality of the system state.Finally,theoretical analysis and experimental results verify that the proposed method is more effective to protect privacy of users.3.The rumor propagation behavior in online social networks studied in this research.The LDA topic model and relative entropy theory used to verify and reveal the potential patterns between the inherent properties of rumors and the dynamic characteristics of dissemination.In addition,in order to describe the propagation behavior of rumors in heterogeneous networks more accurately,we establish a two-layer rumor propagation model based on community clustering by introducing complex network theory and infectious disease dynamics model,which can reflect the characteristic of non-constant spreading rate in complex social networks.Experiments on real data sets show that the novelty of rumors has an important influence on the propagation process,and its propagation behavior does not simply obey the stable exponential growth pattern described by the traditional epidemic model,but has many fluctuations during the process of reaching the finally steady state.