Modeling Human Periodic Behaviors And Its Impact On Worm Propagation

Statistical regularity of human behavior is a scientific system concerned by lots of theorycourses, especially in psychology and economics. Humans are complex aggregations whichare consist of behaviors, so it is a difficult problem to study human behaviors. There are allkinds of resources factors outside affecting human behaviors, at the same time, human'svariety of decision models has an impact on human behaviors in some ways. So the regularityof human behaviors is not easy to be uncovered. Personal behaviors are often used as adriving force which is deeply in the society, technology and economic phenomena. In thissense, it is apparently a big issue of modern science to discover human behaviors.The models of the statistical regularity of human behavior are widely used in variousindustries currently, but they are often described as the Poisson distribution because humanbehaviors happen irregularly and stop from the external influence. In fact, Poisson distributionis used to quantify human activities in the statistical regularity of human behaviors. However,in some recent research about regularity of human behaviors, the non-Poisson features exist inhuman behaviors. To study the existing characteristics and phenomenon in the statisticalregularity of human behavior, it is necessary to capture the data array of users'behaviors. Theresearch of the statistical regularity of human behavior usually use concentrated fixed humanactivities database. So this paper records the interval time of consecutive bursty behaviorsfrom the dataset, and simulates humans' behavior operation to study the bursty feature ofhuman behaviors.Human's everyday behaviors look like complicated, but most of the behaviors aredecided by the regularity of human behavior at ordinary times. The researchers discover thatthe percentage of human behaviors not based on cyclical behaviors is less than five.Psychology's many researches of human suggest that the formation of behavior's regularityneeds to take some certain time. A behavior repeated by human for21days can transform intoa cyclical behavior, and if the repeating time is over90days, the behavior has even alreadybecome fixed. This work is to research that the behavior cycle has important influences on thestatistical regularity of human behavior. Therefore, the model of behavior cycle on thestatistical regularity of human behavior is put forward to be analyzed with the data from themodel and real data set. The security of the computer is a field considered seriously by people, and more andmore researchers come into this area. To prevent from the outbreak of worms on the network,the study of the mechanism in worm's propagation and the actions between the infectors arenecessary. On the basis of model of periodic behaviors, this work is to research on theinfluence of human behavior on the worm propagation process further. With building anetwork model of three thousand nodes, model of periodic behaviors and Poisson modelboth simulate the process of worm spread. And it is to find out the cause of the real reason fordifferent decay time by analyzing the data of different network in scale of two models forinfection and controlling the time interval of new infection. This work further finds out thathuman behavior is a reason causing long prevalence time besides network topologies andvirus infecting styles. The results show that worm spread process has a long stage of a slowstart due to the non-Poisson nature of the human behavior's statistical regularity, and thelength of the stage increases with extending of network scales.