吉林大学学报(理学版) ›› 2024, Vol. 62 ›› Issue (3): 629-635.

• • 上一篇    下一篇

自适应特征融合的迭代实体对齐方法

李婷婷, 邵斐, 温天晓, 董飒   

  1. 吉林大学 计算机科学与技术学院, 符号计算与知识工程教育部重点实验室, 长春 130012
  • 收稿日期:2023-07-12 出版日期:2024-05-26 发布日期:2024-05-26
  • 通讯作者: 董飒 E-mail:dongsa@jlu.edu.cn

Iterative Entity Alignment Method for Adaptive Feature Fusion

LI Tingting, SHAO Fei, WEN Tianxiao, DONG Sa   

  1. Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, College of Computer Science and Technology, Jilin University, Changchun 130012, China
  • Received:2023-07-12 Online:2024-05-26 Published:2024-05-26

摘要: 针对知识图谱实体对齐任务中缺乏训练数据以及长尾实体对齐准确率较低的问题, 提出一种基于自适应特征融合策略的迭代实体对齐方法, 并设计一种迭代策略自动扩充训练数据的规模. 该方法使用知识图谱的结构信息, 并利用关系、 属性和实体名称信息作为语义信息辅助对齐, 从而提升对齐效果. 在数据集上的实验结果表明, 该模型在知识图谱实体对齐任务中效果良好.

关键词: 知识图谱, 实体对齐, 迭代策略, 自适应特征融合

Abstract: Aiming at the problems of insufficient training data and low accuracy of long-tail entity alignment  in the task of knowledge graph entity alignment, we  proposed an iterative entity alignment method based on an adaptive feature fusion strategy and designed an iterative strategy to automatically expand the scale of the training data. This method utilized the structural information of the knowledge graph and utilized  relationships, attributes, and entity name information as  semantic information to assist  alignment 
and  improve alignment effectiveness. The experimental results on the dataset show that the proposed model  performs well in the task of knowledge graph entity alignment.

Key words: knowledge graph, entity alignment, iterative strategy, adaptive feature fusion

中图分类号: 

  • TP391