Research On Influence Evaluation Of Nodes In Social Networks

The development of social network theory provides us with new ways to understand the phenomenon of the world.Most of the systems in the real world,including social systems and biological systems can be modeled as network structures in some way.A collection of individuals together with the social relations forming the network,among which a small portion of nodes is closely related with the structure and play key role on the function of the networks.For the spreading processes in networks,someone may affect other nodes in a larger scope due to their particularity in networks.In this way,the discovery of influential nodes in networks has great theoretical significance and wide practical value.The study for identifying influential nodes based on network structure analysis and spreading dynamics is presented in this dissertation.The importance of nodes relates much to the its neighbors,and many studies have taken nodes' multi-step neighbors and local connection structures into account to measure their influence.In the description of local connections,considering that the clustering coefficient cannot effectively reflect the scale of local connections,this centrality definition leverages the proportion of triangle structures to quantify the structural characteristics among the neighbors of the node.Then,we propose a new measure to rank the influence of nodes,which is called Local Triangle Structure Centrality(LTSC).To evaluate the performance of the proposed LTSC method,we apply it method on both synthetic and real networks.Comparing LTSC with other seven centrality measures in terms of distinguishability and find that LTSC is effective in assignment of distinct ranks to nodes with different spreading capabilities.Further,the SIR model is employed to simulate the real spreading process.By Kendall's tau rank correlation coefficient,we compute the rank correlation between the two ranks generated by the SIR model and one of the centrality measures.The results on real social network and synthetic networks such as the BA mode network and the LFR mode network model demonstrate that ranking result by LTSC method is better correlated with the real spreading process and outperforms the other local and semi-local methods in evaluating the node's influence in most cases.Furthermore,other comprehensive experiments also demonstrate that LTSC is more accurate in the identifying the most influential nodes.As the practical application of the proposed LTSC method,this dissertation collects and collates the cooperation relations of some key members in the field of management science and engineering,and then the scientific collaboration network which contains 3046 researchers and 5062 cooperative links is constructed here.The influence of the scientific research is evaluated by the proposed LTSC here and the influential nodes in the network can be identified easily.The research result may be of great significance for identifying influential nodes by the topological features of networks,by analyzing the influential nodes in the network providing reference for scientific research cooperation and discipline construction.