The model we proposed synthesized the topological characteristics of the network to construct a multiindex evaluation matrix which was converted by Gaussian kernel function, and we got a new comprehensive and balanced multidimensional data by using the eigenvalues contribution as weights coefficient. We eventually obtained a one-dimensional evaluation vector based on the local optimization reconstruction of the data, and we verified the validity of the model using followed simulations. On the basis of the model, by defining a concept which called community element similarity, constructing a ma...
