Modeling For Associated Information Propagate In Online Social Network

In Online Social Networks(OSNs) multiple contagions not only propagate through the network but also interact with each other at the same time. In order to understand the diffusion process of contagions it is necessary to study how different contagions interact. Most of prior work considered individual contagions as independent and thus spreading in isolation. In this paper, an analogy is made between OSNs and biology systems. The interaction among contagions could be regarded as the competing among species. We propose an Interaction Diffusion Model(IDM) which is based on the clas-sic reaction diffusion equation in dynamic biology systems to describe and predict the interactions among multiple contagions. Two real datasets collected from Github are used to testify the predicting performance of the IDM model. Experimental results show that IDM model outperforms the compared models at predicting accuracy.In the online social network (OSNs), all information are not only traveling through the network independently, but also communicating with each other in the process. In order to understand the diffusion process associated social network information, when the social network with a plurality of associated information and communication be-tween each other, if they compete (spread communication, dissemination of informa-tion A can inhibit B information or information B can inhibit A information) or co-operative (spread information A incidentally, there is conducive to the dissemination of information, the B or B information has spread to incidentally information A) rela-tionship, how to through the existing historical data to predict the total amount of each information in the entire internal network in the next period of time is how to change. We will do this social network and has a food chain ecosystem of analogy, the interac-tion between information can be viewed as the competition between species, diffusion model based on dynamic ecosystem classic reaction, the IDM model is established. From the Github and Digg is a collection of two real data sets, and is proved by the experiment results:compared with the Lotka-Volterra model and Fisher model, IDM model has better prediction performance. Our contribution:·Put forward a hypothesis, for the information A and information B, in the same social networks in different sub networks, they are in the process of communi-cation interaction may not be the same.·To study the propagation process of multi information with the relationship be-tween competition and cooperation in a social network by introducing the clas-sical dynamic biological systems of reaction diffusion model.· Through the analysis of parameters of IDM model, we can analyze the interac-tive relationship between information in each sub network, and the utilization efficiency of the dissemination of information that we gain.· We used two real datasets from Github and Digg to validate the accuracy of the IDM model.