基于嵌入特征和稀疏矩阵的实体对齐方法

冯超文,耿程晨,刘英莉

Entity alignment method based on embedding features and sparse matrices

Chaowen FENG,Chengchen GENG,Yingli LIU

表 2 不同实体对齐方法在DBP15K数据集上的性能对比

Tab.2 Performance comparison of different entity alignment methods on DBP15K dataset

方法 DBP_ZH-EN DBP_JA-EN DBP_FR-EN Hits@1 Hits@10 MRR Hits@1 Hits@10 MRR Hits@1 Hits@10 MRR MTransE[16] 0.209 0.512 0.310 0.250 0.572 0.360 0.247 0.577 0.360 GCN-Align[17] 0.434 0.762 0.550 0.427 0.762 0.540 0.411 0.772 0.530 MuGNN[18] 0.494 0.844 0.611 0.501 0.857 0.621 0.495 0.870 0.621 BootEA[19] 0.629 0.847 0.703 0.622 0.853 0.701 0.653 0.874 0.731 PSR[20] 0.802 0.935 0.851 0.803 0.938 0.852 0.828 0.952 0.874 MRAEA[21] 0.757 0.930 0.827 0.758 0.934 0.826 0.781 0.948 0.849 AttrGNN[22] 0.796 0.929 0.845 0.783 0.920 0.834 0.919 0.979 0.910 RDGCN[23] 0.697 0.842 0.750 0.763 0.897 0.810 0.873 0.950 0.901 HGCN[24] 0.720 0.857 0.760 0.766 0.897 0.812 0.892 0.961 0.910 JEANS[25] 0.719 0.895 0.791 0.737 0.914 0.798 0.769 0.940 0.827 EPEA[26] 0.885 0.953 0.911 0.924 0.969 0.942 0.955 0.986 0.967 SNGA[27] 0.987 0.997 0.991 0.991 0.998 0.994 0.998 1.000 0.999 EAMI[28] 0.935 0.982 0.950 0.939 0.978 0.950 0.987 0.996 0.990 本研究 0.871 0.950 0.900 0.938 0.982 0.955 0.976 0.995 0.984