Graph clustering, which aims to divide the nodes in the graph into several distinct clusters, is a fundamental and challenging task. In recent years, deep graph clustering methods have been increasingly proposed and achieved promising performance. However, the corresponding survey paper is scarce and it is imminent to make a summary in this field. From this motivation, this paper makes the first comprehensive survey of deep graph clustering. Firstly, the detailed definition of deep graph clustering and the important baseline methods are introduced. Besides, the taxonomy of deep graph clustering methods is proposed based on four different criteria including graph type, network architecture, learning paradigm, and clustering method. In addition, through the careful analysis of the existing works, the challenges and opportunities from five perspectives are summarized. At last, the applications of deep graph clustering in four domains are presented. It is worth mentioning that a collection of state-of-the-art deep graph clustering methods including papers, codes, and datasets is available on GitHub. We hope this work will serve as a quick guide and help researchers to overcome challenges in this vibrant field.

本文是关于深度图聚类的首个全面综述，首先介绍了深度图聚类的定义和基准方法，然后提出基于四种不同标准的深度图聚类方法的分类学。此外，通过对现有工作的仔细分析，总结了来自五个方面的挑战和机遇。最后，介绍了深度图聚类在四个主要领域中的应用，并提供了 GitHub 上的深度图聚类方法的最新文献和数据集。

深度图聚类的调查: 它的分类，挑战和应用