面向节点分类的图神经网络节点嵌入增强模型
|
|
曾菊香,王平辉,丁益东,兰林,蔡林熹,管晓宏
|
Graph neural network based node embedding enhancement model for node classification
|
|
Ju-xiang ZENG,Ping-hui WANG,Yi-dong DING,Lin LAN,Lin-xi CAI,Xiao-hong GUAN
|
|
| 表 2 Cora数据集上的分类准确度 |
| Tab.2 Prediction accuracy on Cora dataset % |
|
| 方法 | 5节点/类 | 10节点/类 | 20 节点/类 | | Raw | 36.4±3.0 | 40.9±2.5 | 47.1±1.9 | | MLP | 38.5±4.2 | 46.8±3.8 | 54.3±3.1 | | DeepWalk | 53.6±4.5 | 60.7±3.5 | 66.6±1.9 | | ChebyNet | 60.3±6.6 | 70.3±4.5 | 76.9±2.2 | | GCN | 68.2±5.6 | 75.4±2.2 | 79.6±1.6 | | GAT | 72.2±4.5 | 77.5±1.9 | 81.1±1.7 | | GCN-DSM | *69.4±5.3 | *76.2±3.0 | 79.5±1.9 | | GAT-DSM | *75.1±4.1 | *79.2±1.7 | *81.8±1.4 |
|
|
|