Multiplex Graph Clustering with Network Embedding Technique: Review, Methods and Challenges

Abstract

Multiplex graph clustering with network embedding technique has received considerable attention recently. The multiplex graph, a special multilayer network, can model the various interactions among similar entities, and it can describe more comprehensive information than the traditional single-layer network. In this paper, we study the multiplex graph clustering methods with network embedding technique, which aims at learning low-dimensional embedding for each node in the graph. The obtained node representations can be inputted into the clustering algorithm to generate the graph clustering result. We first reviewed the single-layer and multiplex network embedding techniques for the multiplex graph clustering. Then, we present some important multiplex network datasets for evaluating the proposed clustering methods, and describe some common evaluation methods. Finally, the pros and cons of these method, and challenges are discussed.