Modeling And Analysis Of Complex Network Based On Instant Messaging Apps Information Flow Network

With the rapid development of mobile Internet, mobile instant messaging apps(MIM) as the highest usage apps have increasingly important position and role. MIMhas integrated functions such as social behavior, communication, marketing andentertainment. As the multi-functions information service platform, MIM has hugecommercial value. Thus, study of the characteristics and the formation process of MIMinformation flow network has significant meanings.In this paper, WeChat is selected as the research subject. WeChat is called as theship ticket of mobile Internet in China. As the largest mobile Internet platform-levelapplication, research of WeChat information flow network is more useful andmeaningful. The content of this paper is divided into three parts.First, after literature review, evaluation system of MIM information flow networkis put forward based on complex network and information flow model. The evaluationsystem contains CNI indicators that study the structural characteristics of informationflow network and IFI indicators that study the affecting and spreading mechanism ofinformation flow. Based on evaluation system, assumptions of strong-weak relationshipmodel, no direct and no weight WeChat information flow network model, no direct andwith weight WeChat information flow network model in empirical research are putforward.Second, in the empirical study contains three aspects. In the first part, users’behavior, degree distribution and strong-weak relationship model are studied based onthe whole amount of data. Furthermore, this paper compares MIM and IM on PC device.In the second part, the sample data is obtained by using data collection algorithms anddata mining tools based on the whole amount of data. After that, structuralcharacteristics of WeChat information flow network are studied by using social networkanalysis (SNA) based on the evaluation system of MIM information flow network. Inthis paper,337011nodes and353024edges are obtained to compose WeChatinformation flow network. Empirical research about WeChat information flow networkis carried out by analyzing IFI and CNI indicators such as average path length, networkdiameter, node degree distribution, clustering coefficient, eccentricity, closeness centrality and betweenness centrality. In the third part, different information flowcarriers such as MIM, SMS and Weibo are compared.Third, formation model of MIM information flow network is put forward based onempirical study and literature review about formation model of scale-free network. Inthe model aspect, this paper not only considers increase of edges by adding new nodesbut also by connecting old nodes. The formation model in this paper can also adjustbetween the degree preference and random connection when making new edges byconnecting nodes. In the model analysis aspect, stable degree distribution is proved toexist and is deduced by stochastic probability method. In the numerical simulationaspect, quantitative influence of model parameters on the formation of network isstudied and valuation method of model parameters based on empirical network is putforward. In the model simulation aspect, the power law exponent in the formationmodel coincides with that in the empirical study about MIM information flow network.Thus, the formation model proposed in this paper can simulate, control and predict theactual network well.