Abstract
Graph clustering aims to divide the nodes in the graph into several different clusters and is a basic but challenging task. With the powerful representation capabilities of deep learning, deep graph clustering methods have achieved great success in recent years. However, corresponding survey papers are relatively scarce and a summary of this field is necessary. Motivated by this motivation, we conduct a comprehensive survey of depth map clustering
- First, we present the formal definition, evaluation, and development of the field
- Secondly, a taxonomy of deep graph clustering methods is proposed based on four different criteria, including graph type, network architecture, learning paradigm and clustering method.
- Third, we carefully analyze existing methods through extensive experiments and summarize the challenges and opportunities from five perspectives, including graph data quality, stability, scalability, discriminative ability, and number of unknown clusters
- In addition, the application of deep graph clustering method in six fields including computer vision, natural language processing, recommendation system, social network analysis, bioinformatics and medical science is introduced.
- Last but not least, this article provides open source support including
- 1) Collection of state-of-the-art deep graph clustering methods (papers, codes and datasets) (https://github.com/yueliu1999/Awesome-Deep-Graph-Clustering
- 2) Unified framework for depth map clustering (ht