Research For Determining The Important Node And Indentifying The Role Of Node Based On Topology Of Network And Multiple Indicators

In today’s world, there is a wide range of contacts and exchange of information. However, most of them are in the form of complex network. In the course of analyzing these networks, researchers will focus on those compelling primary nodes and the roles of these important nodes. If we have a special significance analysis of these nodes, then we will discover more valuable and more meaningful information in many actual references, such as the key points in information dissemination network and the leader position of animal family network.Based on the above special significance and in the study stage, finding the core of the network and the nodes which are playing the role of passed hub has become a classic problem of social network analysis. It’s also an important point in other interdisciplinary areas beyond complex network, such as Internet search. However, it is not sufficient to measure the multiple and complex roles of nodes in a network based on a single indicator of network topology information. Selecting the multiple indicators whose correlation is not strong is a more intuitive approach. However, the correlation between these topologies is macroeconomic indicators law, with discrepancies in order to show the special value of important nodes.Based on the analysis of the correlation between topological indices and experimental analysis.We select degree centrality, betweenness centrality and eigenvector centrality as an evaluation portfolio which can determine important nodes and indentify the roles of them. However, taking into account the effect of network size and the scope of application of topological indices, so we choice that we can compute betweenness centrality in the ego network of node and compute the index of eigenvector centrality using the two-step self-centered center network of node. Then we add the three indicators as a comprehensive measure of nodes to determine the important nodes. Furthermore, we compare the sort result of degree centrality and the sort results of other indicators of each node,get the deviation between degree and other indicators of each node, and give the decision threshold based on the structural properties of the network, thus form the identification vectors of each node. Thereby we can identify the roles of nodes. Three indicators selected above take node individual and environmental characteristics into account, which are calculated solely rely on ego networks of nodes or two-step self-centered networks of them. They can reflect the importance and the role of node under the global topological properties from the local topology of the node effectively, and they also reduce the computational complexity of determination and identification process effectively. This paper selects a number of different areas of real datasets and a small amount of simulated datasets to do many experiment and the results show that the proposed method are effective and relatively accurate to determine and sort the importance of nodes in the network, and they can also identify the important roles of nodes such as the central node and the bridging nodes in network topology effectively.