Scalable sparse bipartite graph factorization for multi-view clustering

Jinghan Wu, Ben Yang, Shangzong Yang, Xuetao Zhang, Badong Chen

Abstract: Highlights•The proposed bipartite graph factorization framework demonstrates high efficiency.•The introduced L2,log<math><msub is="true"><mrow is="true"><mi is="true">L</mi></mrow><mrow is="true"><mn is="true">2</mn><mo is="true">,</mo><mi is="true">l</mi><mi is="true">o</mi><mi is="true">g</mi></mrow></msub></math>-(pseudo) norm provide the desired level of sparsity for embedding.•An efficient optimization algorithm is designed for the L2,log<math><msub is="true"><mrow is="true"><mi is="true">L</mi></mrow><mrow is="true"><mn is="true">2</mn><mo is="true">,</mo><mi is="true">l</mi><mi is="true">o</mi><mi is="true">g</mi></mrow></msub></math>-(pseudo) norm-regularized problem.•Extensive experiments illustrate the superiority of the proposed method.