In this paper, we investigate how the sparse representa-tion of network topology can be possible with a dictionaryconstructed by a dictionary learning algorithm in sparsemodeling.Sparse modeling is a statistical approach for estimatingunobserved model parameters from a small number of ob-servations using the sparsity of model parameters.Sparse modeling has been applied to several fields such assignal processing, image processing, and compressed sens-ing. By using sparse modeling, unobserved modelparameters that are used to generate observations can be es-timated from a limited number of (generally noisy) obser-vations.
展开▼
