Network Embedding Architecture using Laplace Regularization-Non-Negative Matrix Factorization for Virtualization

Abstract While modeling the applications for a problem in cloud computing, researchers and scientists frequently use graphs as abstractions. Graphs provide structural models that make it possible to analyze and understand how many separate systems act together. The omnipresence in cloud computing systems is increasing information networks. The graph embedding algorithms preserve the microscopic structure over the cloud, and many of them miss the mesoscopic structure of the networks. In this paper, asymmetric non-negative Laplace regularization for cloud platform and matrix factorization is implemented for network embedding. The proposed algorithm preserves the mesoscopic structure in cloud computing, the learned model from the Laplace, and matrix factorization. The embedded cloud network can be used for link prediction, vertex recommendation, node clustering. It is a scalable algorithm for higher proximity preserving along with community structure. The correctness and convergence are measures as performance parameters in the network. Based factorization is used for updating the rules. The experimental study shows that the proposed system is well-organized compared to the existing process in structure preservation in cloud computing.