Convex Nonnegative Matrix Factorization on Bipartite Graph

In this paper, we propose a novel convex nonnegative matrix factorization (CNMF) method to learn a bipartite graph with exactly k connected compo-nents, where k is the number of clusters. The new bipartite graph learned in our model approximates the original graph but maintains an explicit cluster struc-ture, from which we can immediately get the clustering results. Besides, inspired by self-expressive assumption of the data and subspace clustering, we added ad-ditional constraints to further recover global structural information of the data. Multiplicative updating rules are developed. The superiority of our method over the effectiveness and efficiency is demonstrated by extensive experiments.