自适应属性选择的实体对齐方法

被引:7
作者
苏佳林 [1 ,2 ]
王元卓 [1 ]
靳小龙 [1 ]
程学旗 [1 ]
机构
[1] 中国科学院计算技术研究所网络数据科学与技术重点实验室
[2] 中国科学院大学计算机与控制学院
基金
国家重点研发计划;
关键词
知识图谱; 实体对齐; 自适应属性选择; 联合表示学习; 属性强约束;
D O I
暂无
中图分类号
TP391.1 [文字信息处理];
学科分类号
摘要
现有实体对齐方法普遍存在传统方法依赖外部信息和人工构建特征,而基于表示学习的方法忽略了知识图谱中的结构信息的问题。针对上述问题,提出自适应属性选择的实体对齐方法,融合实体的语义和结构信息训练基于两个图谱联合表示学习的实体对齐模型。提出使用基于自适应属性选择的属性强约束模型,根据数据集特征自动生成最优属性类型和权重约束,提升实体对齐效果。两个实际数据集上的试验表明,该方法与传统表示学习方法相比准确率最高提升了约11%。
引用
收藏
页码:14 / 20
页数:7
相关论文
共 30 条
  • [1] Automatic interlinking of music datasets on the semanticweb. Raimond Y,Sutton C,Sandler M. Proceedings of the1st Workshop about Linked Data on the Web . 2008
  • [2] Automatic interlinking of music datasets on the semanticweb. Raimond Y,Sutton C,Sandler M. Proceedings of the1st Workshop about Linked Data on the Web . 2008
  • [3] Entity alignment between knowledge graphs using attribute embeddings. TRSEDYA B D,Qi J,ZHANG R. AAAI Thirty-Third Conference on Artificial Intelligence . 2019
  • [4] Entity alignment between knowledge graphs using attribute embeddings. TRSEDYA B D,Qi J,ZHANG R. AAAI Thirty-Third Conference on Artificial Intelligence . 2019
  • [5] Discovering and Maintaining Links on the Web of Data. Volz J,Bizer C,Gaedke M,et al. Proceedings of the 8th International Semantic Web Conference . 2009
  • [6] Discovering and Maintaining Links on the Web of Data. Volz J,Bizer C,Gaedke M,et al. Proceedings of the 8th International Semantic Web Conference . 2009
  • [7] Limes-a time-efficient approach for large-scale link discoveryon the web of data. Ngomo A C N,Auer S. Integration . 2011
  • [8] Limes-a time-efficient approach for large-scale link discoveryon the web of data. Ngomo A C N,Auer S. Integration . 2011
  • [9] Learning Knowledge Representation Across Knowledge Graphs. Cai P,Li W,Feng Y,et al. AAAI 2017 Workshop on Knowledge-Based Techniques for Problem Solving and Reasoning (KnowProS 17) . 2017
  • [10] Learning Knowledge Representation Across Knowledge Graphs. Cai P,Li W,Feng Y,et al. AAAI 2017 Workshop on Knowledge-Based Techniques for Problem Solving and Reasoning (KnowProS 17) . 2017