Research On Topological Structure And Evolution Model Of Micro Blog Based On Complex Network

With the rapid development of internet technology and the popularization of the smart phones,MicroBlog plays an important role on China's online social networks platform,which attracts wide attention of scholars domestic and overseas.A lot of network in our lives can be described by the complex network.Whereas,there are few research applying complex networks to online social networks.The reasons can be summarized as follows: on the one hand,the network scale is gigantic,on the other hand,the node is too complicated,and the interaction between the user relationships is complex.Therefore,it is a challenge to use the theory of complex network to study the online network.In this thesis,a network model is built based on the “follower and followee” relationships by the interdisciplinary ideas and methods,and the topology structure of the network model is analyzed.Most of complex network models are based on undirected network,but the MicroBlog network is a directed network.Hence,it is practical significance to build a model to simulate the real network evolution process for predicting user behaviors,information dissemination and so on.The main work and innovations are as follows:First,the thesis discusses the complex network models and the evolution of the BA models.On the basis of the directed network,adding the new node as the target node and original node to the network respectively,and considering the increase of directed connect in Internal networks.We improve the BA model,calculate the out-degree and in-degree distribution,and verify that they obey the power law distribution.Secondly,MicroBlog user data is crawled via MicroBlog API based on the MicroBlog SDK,the out-degree distribution,in-degree distribution,clustering coefficient and the average path length and other topological characteristics is caculated.The results show that the out-degree distribution,in-degree distribution both approximately obey the power law distribution;the cluster coefficient is higher than the same size of random networks;In the process of caculating the average path,we use the upper bound of the average path method replace the traditional method,improve the speed of calculating average path greatly.The results show that the network has the characteristic of small word properties.Thirdly,the model is proposed before is simulating,and the results of simulation model and the real MicroBlog network model are compared.We found the model can well simulate the real network's evolution.