Hybrid Clustering based on Content and Connection Structure using Joint Nonnegative Matrix Factorization

TL;DR

This paper introduces a hybrid clustering method that combines content and connection structure using joint nonnegative matrix factorization, improving clustering quality on datasets with both text and graph data.

Contribution

The paper proposes a novel joint NMF and SymNMF optimization framework with an efficient algorithm for hybrid clustering, extending to hypergraph and multi-feature data.

Findings

Hybrid method outperforms standard NMF and SymNMF in clustering quality.

Effective for datasets with both content and connection information.

Applicable to real-world data like citation networks and adaptable to multiple features.

Abstract

We present a hybrid method for latent information discovery on the data sets containing both text content and connection structure based on constrained low rank approximation. The new method jointly optimizes the Nonnegative Matrix Factorization (NMF) objective function for text clustering and the Symmetric NMF (SymNMF) objective function for graph clustering. We propose an effective algorithm for the joint NMF objective function, based on a block coordinate descent (BCD) framework. The proposed hybrid method discovers content associations via latent connections found using SymNMF. The method can also be applied with a natural conversion of the problem when a hypergraph formulation is used or the content is associated with hypergraph edges. Experimental results show that by simultaneously utilizing both content and connection structure, our hybrid method produces higher quality clustering results compared to the other NMF clustering methods that uses content alone (standard NMF) or connection structure alone (SymNMF). We also present some interesting applications to several types of real world data such as citation recommendations of papers. The hybrid method proposed in this paper can also be applied to general data expressed with both feature space vectors and pairwise similarities and can be extended to the case with multiple feature spaces or multiple similarity measures.