Study Of The Dynamic Evolution Of An OSN Structure And The Impacts On Dissemination Of Information

With the development of Web2.0and other Internet technologies, online socialnetworks have been widely used and are becoming more and more important inpeople’s communication. In this paper we studied the dynamic structural evolutionof a large OSN, and the impacts that the changes of the network topology have ondissemination of information in this OSN. This will do favor to the analysis of theevolution of network structure, to its impacts on information dissemination, tonetwork management and network marketing.We have found that the curves of the number of nodes and the number of edgesin this OSN are similar to the first part of the S-shaped growth curve which hasbeen found by many scholars in other OSNs through studying the structureevolution of this OSN, the network scale is extending rapidly at present stage. Wealso find the average shortest distance and the clustering coefficient of this OSNshow non-monotone feature at the early stage of development of this OSN, butgradually show a monotonic change over a period of time.And this OSN keepssmall-world characteristics during its formation. The degree distribution ofpower-law characteristics is not obvious at the beginning, but it is becoming moreand more significant with the network development, however finally it is not astrict power-law distribution, it shows cut-off phenomenon at the end of the degreedistribution curve. Furthermore, the OSN underwent a transition from degreeassortativity characteristic of collaborative networks to degree disassortativitycharacteristic. These findings are meaningful to the exploration of networkstructural evolution of OSNs, and to the changes of the behavior characteristics ofInternet users.Moreover, this paper simulates information dissemination in the networks atdifferent stages of this OSN. The results of simulation indicate that the proportionof the nodes which have received information to the total nodes of the network isupgrading, as the network clustering coefficient decreases and the average degreeincreases with the development of this OSN. And these show the time of finishinginformation dissemination just grow slightly as it is always in line with theprinciple of "six degrees of separation" although the network average shortestdistance is growing with the development of this OSN.