社会网络的数据规模在不断扩大,现存的异常检测算法对复杂社会网络进行检测的效果不理想,提出了一种基于图模块度聚类的异常检测算法(anomaly detection algorithm based on graph modularity clustering,GMC_AD),该算法适用于解决受网...社会网络的数据规模在不断扩大,现存的异常检测算法对复杂社会网络进行检测的效果不理想,提出了一种基于图模块度聚类的异常检测算法(anomaly detection algorithm based on graph modularity clustering,GMC_AD),该算法适用于解决受网络规模以及复杂度的限制导致检测效率不高的问题。GMC_AD算法在分析网络拓扑结构的基础上,通过引入异常节点加权机制和模块度聚类算法进行异常检测。GMC_AD算法主要在三个方面进行改进:a)设计网络中节点演化的量化策略,以此识别具有异常演化行为的节点来得到异常节点集合;b)通过模块度聚类的方法降低网络规模;c)在计算网络波动值的过程中使用加权机制合理考虑异常节点的影响,再通过网络波动值变化来检测异常。基于真实社会网络VAST、EU_E-mail和ENRON进行对比实验,GMC_AD算法准确地检测出异常发生的时段,实验结果显示在事件检测敏感性上提高了50%~82%,在异常检测运行效率上提高了30%~70%。实验结果表明,GMC_AD算法不仅提高了异常检测算法的准确率和敏感性,还提高了异常检测算法的效率。展开更多
Wireless sensor networks are extremely vulnerable to various security threats.The intrusion detection method based on game theory can effectively balance the detection rate and energy consumption of the system.The acc...Wireless sensor networks are extremely vulnerable to various security threats.The intrusion detection method based on game theory can effectively balance the detection rate and energy consumption of the system.The accurate analysis of the attack behavior of malicious sensor nodes can help to configure intrusion detection system,reduce unnecessary system consumption and improve detection efficiency.However,the completely rational assumption of the traditional game model will cause the established model to be inconsistent with the actual attack and defense scenario.In order to formulate a reasonable and effective intrusion detection strategy,we introduce evolutionary game theory to establish an attack evolution game model based on optimal response dynamics,and then analyze the attack behavior of malicious sensor nodes.Theoretical analysis and simulation results show that the evolution trend of attacks is closely related to the number of malicious sensors in the network and the initial state of the strategy,and the attacker can set the initial strategy so that all malicious sensor nodes will eventually launch attacks.Our work is of great significance to guide the development of defense strategies for intrusion detection systems.展开更多
An evolution model of KAD Dynamic Model Network(KDMN) is proposed to study the reason of hot node and simulate the process of network evolution based on node behavior from a holistic perspective.First,some symbols and...An evolution model of KAD Dynamic Model Network(KDMN) is proposed to study the reason of hot node and simulate the process of network evolution based on node behavior from a holistic perspective.First,some symbols and meanings are introduced to describe nodes relationship and network states at a time step.Second,some evolution rules for network are formulated when node behaviors of join,exit,routing table update,data retrieval and content index distribution happen with different contextual scene in KAD network.In addition,a lightweight simulator is designed to implement the KDMN model.Moreover,an example of how to use the simulator to simulate the network changes in order to observe the result is described in detail.Finally,the KDMN is applied to analyze the reason for the formation of hot nodes in the BT and eMule network in the experiment.The different evolution principles of local priority,global priority and hybrid random are adopted based on the provision of network protocol of BT and eMule.The result of this experiment demonstrates that there are some hot nodes exist in the KAD network.However,the principle of hybrid random can effectively alleviate the phenomenon that a node is widely linked with others compared with global and local priority.展开更多
In this paper,we study the dynamic evolution of friendship network in SNS(Social Networking Site).Our analysis suggests that an individual joining a community depends not only on the number of friends he or she has wi...In this paper,we study the dynamic evolution of friendship network in SNS(Social Networking Site).Our analysis suggests that an individual joining a community depends not only on the number of friends he or she has within the community,but also on the friendship network generated by those friends.In addition,we propose a model which is based on two processes:first,connecting nearest neighbors;second,strength driven attachment mechanism.The model reflects two facts:first,in the social network it is a universal phenomenon that two nodes are connected when they have at least one common neighbor;second,new nodes connect more likely to nodes which have larger weights and interactions,a phenomenon called strength driven attachment(also called weight driven attachment).From the simulation results,we find that degree distribution P(k),strength distribution P(s),and degree-strength correlation are all consistent with empirical data.展开更多
Most of existing works on complex network assumed that the nodes and edges were uncapacitated during the evolving process,and displayed "rich club" phenomenon.Here we will show that the "rich club"...Most of existing works on complex network assumed that the nodes and edges were uncapacitated during the evolving process,and displayed "rich club" phenomenon.Here we will show that the "rich club" could be changed to "common rich" if we consider the node capacity.In this paper,we define the node and edge attractive index with node capacity,and propose a new evolving model on the base of BBV model,with evolving simulations of the networks.In the new model,an entering node is linked with an existing node according to the preferential attachment mechanism defined with the attractive index of the existing node.We give the theoretical approximation and simulation solutions.If node capacity is finite,the rich node may not be richer further when the node strength approaches or gets to the node capacity.This is confirmed by analyzing the passenger traffic and routes of Chinese main airports.Due to node strength being function of time t,we can use the theoretical approximation solution to forecast how node strength changes and the time when node strength reaches its maximum value.展开更多
文摘社会网络的数据规模在不断扩大,现存的异常检测算法对复杂社会网络进行检测的效果不理想,提出了一种基于图模块度聚类的异常检测算法(anomaly detection algorithm based on graph modularity clustering,GMC_AD),该算法适用于解决受网络规模以及复杂度的限制导致检测效率不高的问题。GMC_AD算法在分析网络拓扑结构的基础上,通过引入异常节点加权机制和模块度聚类算法进行异常检测。GMC_AD算法主要在三个方面进行改进:a)设计网络中节点演化的量化策略,以此识别具有异常演化行为的节点来得到异常节点集合;b)通过模块度聚类的方法降低网络规模;c)在计算网络波动值的过程中使用加权机制合理考虑异常节点的影响,再通过网络波动值变化来检测异常。基于真实社会网络VAST、EU_E-mail和ENRON进行对比实验,GMC_AD算法准确地检测出异常发生的时段,实验结果显示在事件检测敏感性上提高了50%~82%,在异常检测运行效率上提高了30%~70%。实验结果表明,GMC_AD算法不仅提高了异常检测算法的准确率和敏感性,还提高了异常检测算法的效率。
基金National Natural Science Foundation of China(No.61163009)。
文摘Wireless sensor networks are extremely vulnerable to various security threats.The intrusion detection method based on game theory can effectively balance the detection rate and energy consumption of the system.The accurate analysis of the attack behavior of malicious sensor nodes can help to configure intrusion detection system,reduce unnecessary system consumption and improve detection efficiency.However,the completely rational assumption of the traditional game model will cause the established model to be inconsistent with the actual attack and defense scenario.In order to formulate a reasonable and effective intrusion detection strategy,we introduce evolutionary game theory to establish an attack evolution game model based on optimal response dynamics,and then analyze the attack behavior of malicious sensor nodes.Theoretical analysis and simulation results show that the evolution trend of attacks is closely related to the number of malicious sensors in the network and the initial state of the strategy,and the attacker can set the initial strategy so that all malicious sensor nodes will eventually launch attacks.Our work is of great significance to guide the development of defense strategies for intrusion detection systems.
文摘An evolution model of KAD Dynamic Model Network(KDMN) is proposed to study the reason of hot node and simulate the process of network evolution based on node behavior from a holistic perspective.First,some symbols and meanings are introduced to describe nodes relationship and network states at a time step.Second,some evolution rules for network are formulated when node behaviors of join,exit,routing table update,data retrieval and content index distribution happen with different contextual scene in KAD network.In addition,a lightweight simulator is designed to implement the KDMN model.Moreover,an example of how to use the simulator to simulate the network changes in order to observe the result is described in detail.Finally,the KDMN is applied to analyze the reason for the formation of hot nodes in the BT and eMule network in the experiment.The different evolution principles of local priority,global priority and hybrid random are adopted based on the provision of network protocol of BT and eMule.The result of this experiment demonstrates that there are some hot nodes exist in the KAD network.However,the principle of hybrid random can effectively alleviate the phenomenon that a node is widely linked with others compared with global and local priority.
基金Supported by Program for New Centurty Excellent Talents in University under Grant No. NCET-11-0597the Fundamental Research Funds for the Central Universities under Grant No. 2012RC1002
文摘In this paper,we study the dynamic evolution of friendship network in SNS(Social Networking Site).Our analysis suggests that an individual joining a community depends not only on the number of friends he or she has within the community,but also on the friendship network generated by those friends.In addition,we propose a model which is based on two processes:first,connecting nearest neighbors;second,strength driven attachment mechanism.The model reflects two facts:first,in the social network it is a universal phenomenon that two nodes are connected when they have at least one common neighbor;second,new nodes connect more likely to nodes which have larger weights and interactions,a phenomenon called strength driven attachment(also called weight driven attachment).From the simulation results,we find that degree distribution P(k),strength distribution P(s),and degree-strength correlation are all consistent with empirical data.
基金supported by the National Natural Science Foundation of China (Grant Nos.71171111,70771046 and 71201081)the Colleges Graduate Research and Innovation Program of Jiangsu Province (Grant Nos.CXZZ11-0220 and CX10B-102Z)
文摘Most of existing works on complex network assumed that the nodes and edges were uncapacitated during the evolving process,and displayed "rich club" phenomenon.Here we will show that the "rich club" could be changed to "common rich" if we consider the node capacity.In this paper,we define the node and edge attractive index with node capacity,and propose a new evolving model on the base of BBV model,with evolving simulations of the networks.In the new model,an entering node is linked with an existing node according to the preferential attachment mechanism defined with the attractive index of the existing node.We give the theoretical approximation and simulation solutions.If node capacity is finite,the rich node may not be richer further when the node strength approaches or gets to the node capacity.This is confirmed by analyzing the passenger traffic and routes of Chinese main airports.Due to node strength being function of time t,we can use the theoretical approximation solution to forecast how node strength changes and the time when node strength reaches its maximum value.
