四川轻化工大学学报（自然科学版）2024，Vol.37Issue(4)：47-57,11.DOI:10.11863/j.suse.2024.04.06
深度聚类研究综述
A Review of Deep Clustering Studies
摘要
Abstract
Clustering plays an important role in machine learning,and learning a good representation of data is essential for clustering.Deep clustering joint optimization representation learning and clustering models have been widely used in various clustering tasks.The paper overviews the recent advances in deep clustering and summarize its applications in different fields.Firstly,the basic concepts and principles of deep clustering are introduced,and from the perspective of neural network model,it is divided into autoencoder-based,variational-autoencoder-based,generative adversarial network-based,twin-based network/contrastive learning,and deep clustering based on graph neural network,which are analyzed and summarized respectively.Then,the applications of deep clustering in the fields of image,text,and recognition detection are discussed.Finally,the current hotspots and future development directions of deep clustering research are prospected.关键词
深度学习/神经网络/特征提取/聚类Key words
deep clustering/neural networks/feature extraction/clustering分类
信息技术与安全科学引用本文复制引用
黄雪艳,张翠红,赵薇,王继奎..深度聚类研究综述[J].四川轻化工大学学报（自然科学版）,2024,37(4):47-57,11.基金项目
国家自然科学基金项目(12201267) （12201267）
