Research Of Microblog Information Diffusion Network Properties

Microblog is a milestone in the field of Internet and is only developed rapidly in recent years. Microblog has completely changed people’s daily life and behavior. Because of the simple operation and function of practical advantages, it attracted a huge number of users since the birth of microblog. Research of microblogg properties can be studied from user relation network, information diffusion networks and diffusion mechanisms three aspects. Users are the subject and core of microblog. The ’follow’ and ’be followed’ relation between users has formed a network of relationships. Obviously, this is a directed network. The network is the carrier of information diffusion, and the spreading the message of each information post require the help of this network.Firstly, this paper introduces some basic knowledge of microblog, including the concept of microblog, the current research, development, main applications and key research issues. Microblog is a complex system, which contains a variety of different means of special symbols. It also reflected in the complexity of users’diversity, information diversity and the use motivation of microblog diversity. In this paper, the users and information are classified. Information has been able to diffuse rapidly in the microblog system mainly rely on its own transmission power. Therefore, it is necessary to study the transition power of microblog to reveal deep-seated microblog feature.Relations between microblog users form a static network diagram, so we can try to study the microblog information diffusion network from the perspective of complex network research. A huge network is built based on the data of users’ relationship. We used VC++, Gephi, MALAB and other tools analyzed the eigenvalues of the network, the clustering coefficient, average path length, K-core and communities. The result show that microblog information diffusion network has small world and scale-free characteristic, and its in-degree and K-core distribution conformed to power-law distribution, which provides a theoretical basis for the construction and improvement of microblog relation model. In the correlation analysis, we found there is a high correlation between eigenvector centrality and clustering coefficient. They are the measure of network centrality indicators. It indicates that a great degree of a network is often with a high diagram density, but the centrality may not strong. In other words, the small world phenomenon is not obvious. Three motif structures are proposed on the information diffusion network. We find that information scattering structure has the largest number from the experiment. And with the increase of information chain, the number is gradually increasing.SI model, SIS model and SIR model are classical models in the fields of disease diffusion research. They once had played a decisive role in the control of some epidemic. These three models are mainly for the disease, but the diffusion of information on microblog is different, such as the spread of diseases is on the impact of weather, climate and other environmental factors. However, the diffusion of information on microblog is affected by users’ mood and living conditions, thus a new diffusion model in line with the microblog network is constructed. It is showed that the model can reveal information diffusion microblog process more accurate by simulating different numerical simulation of four groups.