Empirical Research Of A Complex Network Model On Scientific Collaboration

In recent years, the study of social networks has attracted more and moreattention, and the research of scientific collaboration complex network has alsobecome an increasingly important issue. Discipline construction has always been thefundamental construction of the school, and the discipline Project cooperation can usethe network in the form of the expression.The paper selects a more closely and representative of the subject as the object ofstudy in Jinan University, we construct the interdisciplinary cooperation networkmodel in the weighted condition through the acquisition of empirical data, and usenetwork optimization knowledge and community mining algorithm to model of thenetwork characteristic and quantitative analysis.Through theoretical analysis and simulation, we can draw the followingconclusions. It has24connected sub-networks. The largest connected collaborationsub-network appears to be a scale-free network, and it behaves some like asmall-world. Several different mining algorithm of community structure was analyzedand compared. We find the key nodes in the network through the article number,degree, the betweenness centrality and characteristic factor scores and other indicators.Finally, the robustness to network model were studied. Finally according to theconcept of global efficiency, the network vulnerable points are identified. Randomattacking has a little effect on the network. The efficiency doesn’t change drasticallywith some nodes deleted. But deliberate attacking has a large effect on the networkThe network efficiency is0when delete less than30%large degree or highbetweenness nodes.