西华大学学报（自然科学版）2025，Vol.44Issue(6)：91-98,8.DOI:10.12198/j.issn.1673-159X.5199
基于异构图自编码器的端到端聚类方法
An End-To-End Heterogeneous Graph Clustering Method
夏鹤珑 1李弈霄2
作者信息
- 1. 重庆邮电大学软件学院,重庆 400065
- 2. 西华大学计算机与软件工程学院,四川 成都 610039
- 折叠
摘要
Abstract
Heterogeneous graph clustering is a fundamental and difficult task in data mining.It is a great challenge to complete the clustering process while preserving the structural information of heterogen-eous graphs.Therefore,an end-to-end heterogeneous graph clustering method is proposed,which aims to jointly and optimally learn the heterogeneous graph node representation process and the clustering process.Specifically,we use heterogeneous graph auto-encoders to model heterogeneous graphs and learn their node representations.Meanwhile we jointly guiding the generation of node representations by constructing auxil-iary distributions oriented to clustering.So,the learned node representations not only preserve the structur-al information of heterogeneous graphs,but also make them separate in the vector space for clustering pur-poses.The experimental results show that the joint learning of node representations and clustering has bet-ter performance than the traditional separate learning method.关键词
异构图/图神经网络/聚类/深度学习Key words
heterogeneous graph/graph neural network/clustering/deep learning分类
信息技术与安全科学引用本文复制引用
夏鹤珑,李弈霄..基于异构图自编码器的端到端聚类方法[J].西华大学学报（自然科学版）,2025,44(6):91-98,8.
