数据集为cora数据集，cora数据集由机器学习论文组成，共以下7类：
- 基于案例
- 遗传算法
- 神经网络
- 概率方法
- 强化学习
- 规则学习
- 理论
由cora.content和cora.cities文件构成。共2708个样本，每个样本的特征维度是1433。
下载地址：https://linqs.soe.ucsc.edu/data
cora.content：
每一行由论文id+特征向量+标签构成。
31336 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 Neural_Networks 1061127 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 Rule_Learning 1106406 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Reinforcement_Learning 13195 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Reinforcement_Learning ......
cora.cities：
引用关系：被引论文编号以及引论文编号
35 1033 35 103482 35 103515 35 1050679 35 1103960 35 1103985 35 1109199 35 1112911 ......
读取数据集：
import numpy as np import scipy.sparse as sp import torch from sklearn.preprocessing import LabelBinarizer def normalize_adj(adjacency): adjacency += sp.eye(adjacency.shape[0]) degree = np.array(adjacency.sum(1)) d_hat = sp.diags(np.power(degree, -0.5).flatten()) return d_hat.dot(adjacency).dot(d_hat).tocoo() def normalize_features(features): return features / features.sum(1) def load_data(path="/content/drive/My Drive/nlpdata/cora/", dataset="cora"): """Load citation network dataset (cora only for now)""" print('Loading {} dataset...'.format(dataset)) idx_features_labels = np.genfromtxt("{}{}.content".format(path,dataset), dtype=np.dtype(str)) features = sp.csr_matrix(idx_features_labels[:, 1:-1], dtype=np.float32) encode_onehot = LabelBinarizer() labels = encode_onehot.fit_transform(idx_features_labels[:, -1]) # build graph idx = np.array(idx_features_labels[:, 0], dtype=np.int32) idx_map = {j: i for i, j in enumerate(idx)} edges_unordered = np.genfromtxt("{}{}.cites".format(path, dataset), dtype=np.int32) edges = np.array(list(map(idx_map.get, edges_unordered.flatten())), dtype=np.int32).reshape(edges_unordered.shape) adj = sp.coo_matrix((np.ones(edges.shape[0]), (edges[:, 0], edges[:, 1])), shape=(labels.shape[0], labels.shape[0]), dtype=np.float32) features = normalize_features(features) adj = normalize_adj(adj) idx_train = range(140) idx_val = range(200, 500) idx_test = range(500, 1500) features = torch.FloatTensor(np.array(features)) labels = torch.LongTensor(np.where(labels)[1]) num_nodes = features.shape[0] train_mask = np.zeros(num_nodes, dtype=np.bool) val_mask = np.zeros(num_nodes, dtype=np.bool) test_mask = np.zeros(num_nodes, dtype=np.bool) train_mask[idx_train] = True val_mask[idx_val] = True test_mask[idx_test] = True return adj, features, labels, train_mask, val_mask, test_mask
构建网络：
import numpy as np import scipy.sparse as sp import torch import torch.nn as nn import torch.nn.functional as F import torch.nn.init as init import torch.optim as optim import matplotlib.pyplot as plt from load_cora import * import sys sys.path.append("/content/drive/My Drive/nlpdata/cora/") class GraphConvolution(nn.Module): def __init__(self, input_dim, output_dim, use_bias=True): """图卷积：L*X* heta Args: ---------- input_dim: int 节点输入特征的维度 output_dim: int 输出特征维度 use_bias : bool, optional 是否使用偏置 """ super(GraphConvolution, self).__init__() self.input_dim = input_dim self.output_dim = output_dim self.use_bias = use_bias self.weight = nn.Parameter(torch.Tensor(input_dim, output_dim)) if self.use_bias: self.bias = nn.Parameter(torch.Tensor(output_dim)) else: self.register_parameter('bias', None) self.reset_parameters() def reset_parameters(self): init.kaiming_uniform_(self.weight) if self.use_bias: init.zeros_(self.bias) def forward(self, adjacency, input_feature): """邻接矩阵是稀疏矩阵，因此在计算时使用稀疏矩阵乘法 Args: ------- adjacency: torch.sparse.FloatTensor 邻接矩阵 input_feature: torch.Tensor 输入特征 """ device = "cuda" if torch.cuda.is_available() else "cpu" support = torch.mm(input_feature, self.weight.to(device)) output = torch.sparse.mm(adjacency, support) if self.use_bias: output += self.bias.to(device) return output def __repr__(self): return self.__class__.__name__ + ' (' + str(self.in_features) + ' -> ' + str(self.out_features) + ')' # ## 模型定义 class GcnNet(nn.Module): """ 定义一个包含两层GraphConvolution的模型 """ def __init__(self, input_dim=1433): super(GcnNet, self).__init__() self.gcn1 = GraphConvolution(input_dim, 16) self.gcn2 = GraphConvolution(16, 7) def forward(self, adjacency, feature): h = F.relu(self.gcn1(adjacency, feature)) logits = self.gcn2(adjacency, h) return logits
进行训练和测试：
# ## 模型训练 # 超参数定义 learning_rate = 0.1 weight_decay = 5e-4 epochs = 200 # 模型定义：Model, Loss, Optimizer device = "cuda" if torch.cuda.is_available() else "cpu" model = GcnNet().to(device) criterion = nn.CrossEntropyLoss().to(device) optimizer = optim.Adam(model.parameters(), lr=learning_rate, weight_decay=weight_decay) adjacency, features, labels, train_mask, val_mask, test_mask= load_data() tensor_x = features.to(device) tensor_y = labels.to(device) tensor_train_mask = torch.from_numpy(train_mask).to(device) tensor_val_mask = torch.from_numpy(val_mask).to(device) tensor_test_mask = torch.from_numpy(test_mask).to(device) indices = torch.from_numpy(np.asarray([adjacency.row, adjacency.col]).astype('int64')).long() values = torch.from_numpy(adjacency.data.astype(np.float32)) tensor_adjacency = torch.sparse.FloatTensor(indices, values, (2708, 2708)).to(device) # 训练主体函数 def train(): loss_history = [] val_acc_history = [] model.train() train_y = tensor_y[tensor_train_mask] for epoch in range(epochs): logits = model(tensor_adjacency, tensor_x) # 前向传播 train_mask_logits = logits[tensor_train_mask] # 只选择训练节点进行监督 loss = criterion(train_mask_logits, train_y) # 计算损失值 optimizer.zero_grad() loss.backward() # 反向传播计算参数的梯度 optimizer.step() # 使用优化方法进行梯度更新 train_acc, _, _ = test(tensor_train_mask) # 计算当前模型训练集上的准确率 val_acc, _, _ = test(tensor_val_mask) # 计算当前模型在验证集上的准确率 # 记录训练过程中损失值和准确率的变化，用于画图 loss_history.append(loss.item()) val_acc_history.append(val_acc.item()) print("Epoch {:03d}: Loss {:.4f}, TrainAcc {:.4}, ValAcc {:.4f}".format( epoch, loss.item(), train_acc.item(), val_acc.item())) return loss_history, val_acc_history # 测试函数 def test(mask): model.eval() with torch.no_grad(): logits = model(tensor_adjacency, tensor_x) test_mask_logits = logits[mask] predict_y = test_mask_logits.max(1)[1] accuarcy = torch.eq(predict_y, tensor_y[mask]).float().mean() return accuarcy, test_mask_logits.cpu().numpy(), tensor_y[mask].cpu().numpy() if __name__ == "__main__": train() test_accuracy, _, _ = test(tensor_test_mask) print("测试准确率是：{:.4f}".format(test_accuracy))
结果：
Loading cora dataset... Epoch 000: Loss 1.9681, TrainAcc 0.3286, ValAcc 0.3467 Epoch 001: Loss 1.7307, TrainAcc 0.4786, ValAcc 0.4033 Epoch 002: Loss 1.5521, TrainAcc 0.5214, ValAcc 0.4033 Epoch 003: Loss 1.3685, TrainAcc 0.6143, ValAcc 0.5100 Epoch 004: Loss 1.1594, TrainAcc 0.75, ValAcc 0.5767 Epoch 005: Loss 0.9785, TrainAcc 0.7857, ValAcc 0.5900 Epoch 006: Loss 0.8226, TrainAcc 0.8286, ValAcc 0.5867 Epoch 007: Loss 0.6849, TrainAcc 0.8929, ValAcc 0.6200 Epoch 008: Loss 0.5448, TrainAcc 0.9429, ValAcc 0.6433 Epoch 009: Loss 0.4152, TrainAcc 0.9429, ValAcc 0.6667 Epoch 010: Loss 0.3221, TrainAcc 0.9857, ValAcc 0.6767 Epoch 011: Loss 0.2547, TrainAcc 1.0, ValAcc 0.7033 Epoch 012: Loss 0.1979, TrainAcc 1.0, ValAcc 0.7167 Epoch 013: Loss 0.1536, TrainAcc 1.0, ValAcc 0.7000 Epoch 014: Loss 0.1276, TrainAcc 1.0, ValAcc 0.6700 Epoch 015: Loss 0.1122, TrainAcc 1.0, ValAcc 0.6867 Epoch 016: Loss 0.0979, TrainAcc 1.0, ValAcc 0.6800 Epoch 017: Loss 0.0876, TrainAcc 1.0, ValAcc 0.6700 Epoch 018: Loss 0.0821, TrainAcc 1.0, ValAcc 0.6667 Epoch 019: Loss 0.0799, TrainAcc 1.0, ValAcc 0.6800 Epoch 020: Loss 0.0804, TrainAcc 1.0, ValAcc 0.6933 Epoch 021: Loss 0.0852, TrainAcc 1.0, ValAcc 0.6833 Epoch 022: Loss 0.0904, TrainAcc 1.0, ValAcc 0.6700 Epoch 023: Loss 0.0914, TrainAcc 1.0, ValAcc 0.6700 Epoch 024: Loss 0.0926, TrainAcc 1.0, ValAcc 0.6400 Epoch 025: Loss 0.0953, TrainAcc 1.0, ValAcc 0.6300 Epoch 026: Loss 0.0931, TrainAcc 1.0, ValAcc 0.6467 Epoch 027: Loss 0.0880, TrainAcc 1.0, ValAcc 0.6600 Epoch 028: Loss 0.0851, TrainAcc 1.0, ValAcc 0.6567 Epoch 029: Loss 0.0814, TrainAcc 1.0, ValAcc 0.6600 Epoch 030: Loss 0.0756, TrainAcc 1.0, ValAcc 0.6433 Epoch 031: Loss 0.0709, TrainAcc 1.0, ValAcc 0.6567 Epoch 032: Loss 0.0682, TrainAcc 1.0, ValAcc 0.6467 Epoch 033: Loss 0.0656, TrainAcc 1.0, ValAcc 0.6700 Epoch 034: Loss 0.0628, TrainAcc 1.0, ValAcc 0.6633 Epoch 035: Loss 0.0618, TrainAcc 1.0, ValAcc 0.6833 Epoch 036: Loss 0.0617, TrainAcc 1.0, ValAcc 0.6700 Epoch 037: Loss 0.0615, TrainAcc 1.0, ValAcc 0.6667 Epoch 038: Loss 0.0615, TrainAcc 1.0, ValAcc 0.6600 Epoch 039: Loss 0.0616, TrainAcc 1.0, ValAcc 0.6733 Epoch 040: Loss 0.0613, TrainAcc 1.0, ValAcc 0.6800 Epoch 041: Loss 0.0610, TrainAcc 1.0, ValAcc 0.6867 Epoch 042: Loss 0.0603, TrainAcc 1.0, ValAcc 0.6800 Epoch 043: Loss 0.0598, TrainAcc 1.0, ValAcc 0.6667 Epoch 044: Loss 0.0592, TrainAcc 1.0, ValAcc 0.6833 Epoch 045: Loss 0.0584, TrainAcc 1.0, ValAcc 0.6833 Epoch 046: Loss 0.0575, TrainAcc 1.0, ValAcc 0.6967 Epoch 047: Loss 0.0570, TrainAcc 1.0, ValAcc 0.6900 Epoch 048: Loss 0.0565, TrainAcc 1.0, ValAcc 0.6933 Epoch 049: Loss 0.0560, TrainAcc 1.0, ValAcc 0.6867 Epoch 050: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6900 Epoch 051: Loss 0.0557, TrainAcc 1.0, ValAcc 0.6900 Epoch 052: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6967 Epoch 053: Loss 0.0554, TrainAcc 1.0, ValAcc 0.6867 Epoch 054: Loss 0.0552, TrainAcc 1.0, ValAcc 0.6867 Epoch 055: Loss 0.0550, TrainAcc 1.0, ValAcc 0.6933 Epoch 056: Loss 0.0549, TrainAcc 1.0, ValAcc 0.7000 Epoch 057: Loss 0.0548, TrainAcc 1.0, ValAcc 0.7000 Epoch 058: Loss 0.0547, TrainAcc 1.0, ValAcc 0.7067 Epoch 059: Loss 0.0546, TrainAcc 1.0, ValAcc 0.7000 Epoch 060: Loss 0.0545, TrainAcc 1.0, ValAcc 0.6967 Epoch 061: Loss 0.0545, TrainAcc 1.0, ValAcc 0.6967 Epoch 062: Loss 0.0544, TrainAcc 1.0, ValAcc 0.7067 Epoch 063: Loss 0.0544, TrainAcc 1.0, ValAcc 0.7067 Epoch 064: Loss 0.0543, TrainAcc 1.0, ValAcc 0.7033 Epoch 065: Loss 0.0542, TrainAcc 1.0, ValAcc 0.7000 Epoch 066: Loss 0.0542, TrainAcc 1.0, ValAcc 0.7000 Epoch 067: Loss 0.0542, TrainAcc 1.0, ValAcc 0.7033 Epoch 068: Loss 0.0542, TrainAcc 1.0, ValAcc 0.7067 Epoch 069: Loss 0.0542, TrainAcc 1.0, ValAcc 0.7033 Epoch 070: Loss 0.0543, TrainAcc 1.0, ValAcc 0.7033 Epoch 071: Loss 0.0543, TrainAcc 1.0, ValAcc 0.7000 Epoch 072: Loss 0.0543, TrainAcc 1.0, ValAcc 0.7033 Epoch 073: Loss 0.0544, TrainAcc 1.0, ValAcc 0.7067 Epoch 074: Loss 0.0544, TrainAcc 1.0, ValAcc 0.7067 Epoch 075: Loss 0.0545, TrainAcc 1.0, ValAcc 0.7133 Epoch 076: Loss 0.0545, TrainAcc 1.0, ValAcc 0.7100 Epoch 077: Loss 0.0546, TrainAcc 1.0, ValAcc 0.7133 Epoch 078: Loss 0.0546, TrainAcc 1.0, ValAcc 0.7067 Epoch 079: Loss 0.0547, TrainAcc 1.0, ValAcc 0.7133 Epoch 080: Loss 0.0547, TrainAcc 1.0, ValAcc 0.7033 Epoch 081: Loss 0.0548, TrainAcc 1.0, ValAcc 0.7100 Epoch 082: Loss 0.0549, TrainAcc 1.0, ValAcc 0.7067 Epoch 083: Loss 0.0549, TrainAcc 1.0, ValAcc 0.7133 Epoch 084: Loss 0.0549, TrainAcc 1.0, ValAcc 0.7067 Epoch 085: Loss 0.0550, TrainAcc 1.0, ValAcc 0.7100 Epoch 086: Loss 0.0550, TrainAcc 1.0, ValAcc 0.7033 Epoch 087: Loss 0.0551, TrainAcc 1.0, ValAcc 0.7133 Epoch 088: Loss 0.0551, TrainAcc 1.0, ValAcc 0.7067 Epoch 089: Loss 0.0552, TrainAcc 1.0, ValAcc 0.7100 Epoch 090: Loss 0.0553, TrainAcc 1.0, ValAcc 0.6967 Epoch 091: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7067 Epoch 092: Loss 0.0554, TrainAcc 1.0, ValAcc 0.6900 Epoch 093: Loss 0.0556, TrainAcc 1.0, ValAcc 0.7100 Epoch 094: Loss 0.0557, TrainAcc 1.0, ValAcc 0.6833 Epoch 095: Loss 0.0561, TrainAcc 1.0, ValAcc 0.7033 Epoch 096: Loss 0.0558, TrainAcc 1.0, ValAcc 0.6833 Epoch 097: Loss 0.0557, TrainAcc 1.0, ValAcc 0.7100 Epoch 098: Loss 0.0547, TrainAcc 1.0, ValAcc 0.7133 Epoch 099: Loss 0.0546, TrainAcc 1.0, ValAcc 0.6900 Epoch 100: Loss 0.0555, TrainAcc 1.0, ValAcc 0.7033 Epoch 101: Loss 0.0561, TrainAcc 1.0, ValAcc 0.6700 Epoch 102: Loss 0.0579, TrainAcc 1.0, ValAcc 0.6967 Epoch 103: Loss 0.0577, TrainAcc 1.0, ValAcc 0.6633 Epoch 104: Loss 0.0600, TrainAcc 1.0, ValAcc 0.7000 Epoch 105: Loss 0.0550, TrainAcc 1.0, ValAcc 0.6967 Epoch 106: Loss 0.0540, TrainAcc 1.0, ValAcc 0.6767 Epoch 107: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6967 Epoch 108: Loss 0.0528, TrainAcc 1.0, ValAcc 0.7000 Epoch 109: Loss 0.0571, TrainAcc 1.0, ValAcc 0.6700 Epoch 110: Loss 0.0643, TrainAcc 1.0, ValAcc 0.6933 Epoch 111: Loss 0.0583, TrainAcc 1.0, ValAcc 0.6800 Epoch 112: Loss 0.0533, TrainAcc 1.0, ValAcc 0.6700 Epoch 113: Loss 0.0552, TrainAcc 1.0, ValAcc 0.7067 Epoch 114: Loss 0.0534, TrainAcc 1.0, ValAcc 0.6967 Epoch 115: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6833 Epoch 116: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6800 Epoch 117: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6933 Epoch 118: Loss 0.0601, TrainAcc 1.0, ValAcc 0.6700 Epoch 119: Loss 0.0707, TrainAcc 1.0, ValAcc 0.6667 Epoch 120: Loss 0.0670, TrainAcc 1.0, ValAcc 0.6500 Epoch 121: Loss 0.0574, TrainAcc 1.0, ValAcc 0.6467 Epoch 122: Loss 0.0589, TrainAcc 1.0, ValAcc 0.7033 Epoch 123: Loss 0.0493, TrainAcc 1.0, ValAcc 0.6800 Epoch 124: Loss 0.0591, TrainAcc 1.0, ValAcc 0.6900 Epoch 125: Loss 0.0482, TrainAcc 1.0, ValAcc 0.6600 Epoch 126: Loss 0.0562, TrainAcc 1.0, ValAcc 0.6667 Epoch 127: Loss 0.0538, TrainAcc 1.0, ValAcc 0.6900 Epoch 128: Loss 0.0579, TrainAcc 1.0, ValAcc 0.6867 Epoch 129: Loss 0.0557, TrainAcc 1.0, ValAcc 0.6833 Epoch 130: Loss 0.0615, TrainAcc 1.0, ValAcc 0.6733 Epoch 131: Loss 0.0570, TrainAcc 1.0, ValAcc 0.6667 Epoch 132: Loss 0.0612, TrainAcc 1.0, ValAcc 0.6700 Epoch 133: Loss 0.0669, TrainAcc 1.0, ValAcc 0.6967 Epoch 134: Loss 0.0544, TrainAcc 1.0, ValAcc 0.6767 Epoch 135: Loss 0.0605, TrainAcc 1.0, ValAcc 0.6567 Epoch 136: Loss 0.0546, TrainAcc 1.0, ValAcc 0.6567 Epoch 137: Loss 0.0586, TrainAcc 1.0, ValAcc 0.7033 Epoch 138: Loss 0.0501, TrainAcc 1.0, ValAcc 0.6833 Epoch 139: Loss 0.0600, TrainAcc 1.0, ValAcc 0.7067 Epoch 140: Loss 0.0513, TrainAcc 1.0, ValAcc 0.6633 Epoch 141: Loss 0.0587, TrainAcc 1.0, ValAcc 0.6733 Epoch 142: Loss 0.0556, TrainAcc 1.0, ValAcc 0.6833 Epoch 143: Loss 0.0586, TrainAcc 1.0, ValAcc 0.6967 Epoch 144: Loss 0.0565, TrainAcc 1.0, ValAcc 0.6900 Epoch 145: Loss 0.0586, TrainAcc 1.0, ValAcc 0.6833 Epoch 146: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6767 Epoch 147: Loss 0.0589, TrainAcc 1.0, ValAcc 0.6800 Epoch 148: Loss 0.0562, TrainAcc 1.0, ValAcc 0.6900 Epoch 149: Loss 0.0560, TrainAcc 1.0, ValAcc 0.6933 Epoch 150: Loss 0.0565, TrainAcc 1.0, ValAcc 0.6833 Epoch 151: Loss 0.0547, TrainAcc 1.0, ValAcc 0.6767 Epoch 152: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6967 Epoch 153: Loss 0.0549, TrainAcc 1.0, ValAcc 0.6933 Epoch 154: Loss 0.0567, TrainAcc 1.0, ValAcc 0.7000 Epoch 155: Loss 0.0556, TrainAcc 1.0, ValAcc 0.6867 Epoch 156: Loss 0.0568, TrainAcc 1.0, ValAcc 0.6967 Epoch 157: Loss 0.0558, TrainAcc 1.0, ValAcc 0.6967 Epoch 158: Loss 0.0568, TrainAcc 1.0, ValAcc 0.6867 Epoch 159: Loss 0.0560, TrainAcc 1.0, ValAcc 0.6867 Epoch 160: Loss 0.0563, TrainAcc 1.0, ValAcc 0.6933 Epoch 161: Loss 0.0562, TrainAcc 1.0, ValAcc 0.6967 Epoch 162: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6833 Epoch 163: Loss 0.0556, TrainAcc 1.0, ValAcc 0.7000 Epoch 164: Loss 0.0554, TrainAcc 1.0, ValAcc 0.7033 Epoch 165: Loss 0.0556, TrainAcc 1.0, ValAcc 0.6967 Epoch 166: Loss 0.0553, TrainAcc 1.0, ValAcc 0.6900 Epoch 167: Loss 0.0555, TrainAcc 1.0, ValAcc 0.7033 Epoch 168: Loss 0.0554, TrainAcc 1.0, ValAcc 0.6967 Epoch 169: Loss 0.0560, TrainAcc 1.0, ValAcc 0.6833 Epoch 170: Loss 0.0558, TrainAcc 1.0, ValAcc 0.6900 Epoch 171: Loss 0.0561, TrainAcc 1.0, ValAcc 0.7033 Epoch 172: Loss 0.0560, TrainAcc 1.0, ValAcc 0.6967 Epoch 173: Loss 0.0559, TrainAcc 1.0, ValAcc 0.6833 Epoch 174: Loss 0.0557, TrainAcc 1.0, ValAcc 0.6967 Epoch 175: Loss 0.0554, TrainAcc 1.0, ValAcc 0.7033 Epoch 176: Loss 0.0554, TrainAcc 1.0, ValAcc 0.7033 Epoch 177: Loss 0.0551, TrainAcc 1.0, ValAcc 0.6933 Epoch 178: Loss 0.0553, TrainAcc 1.0, ValAcc 0.6967 Epoch 179: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7000 Epoch 180: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6900 Epoch 181: Loss 0.0556, TrainAcc 1.0, ValAcc 0.7000 Epoch 182: Loss 0.0557, TrainAcc 1.0, ValAcc 0.7033 Epoch 183: Loss 0.0558, TrainAcc 1.0, ValAcc 0.6933 Epoch 184: Loss 0.0556, TrainAcc 1.0, ValAcc 0.6867 Epoch 185: Loss 0.0555, TrainAcc 1.0, ValAcc 0.7033 Epoch 186: Loss 0.0554, TrainAcc 1.0, ValAcc 0.7000 Epoch 187: Loss 0.0552, TrainAcc 1.0, ValAcc 0.6933 Epoch 188: Loss 0.0552, TrainAcc 1.0, ValAcc 0.6933 Epoch 189: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7000 Epoch 190: Loss 0.0554, TrainAcc 1.0, ValAcc 0.7000 Epoch 191: Loss 0.0554, TrainAcc 1.0, ValAcc 0.6933 Epoch 192: Loss 0.0555, TrainAcc 1.0, ValAcc 0.7000 Epoch 193: Loss 0.0555, TrainAcc 1.0, ValAcc 0.7000 Epoch 194: Loss 0.0555, TrainAcc 1.0, ValAcc 0.6933 Epoch 195: Loss 0.0554, TrainAcc 1.0, ValAcc 0.6933 Epoch 196: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7000 Epoch 197: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7033 Epoch 198: Loss 0.0553, TrainAcc 1.0, ValAcc 0.6933 Epoch 199: Loss 0.0553, TrainAcc 1.0, ValAcc 0.7033 测试准确率是：0.6480
最后是进行tsne降维可视化：
from sklearn.manifold import TSNE test_accuracy, test_data, test_labels = test(tensor_test_mask) tsne = TSNE(perplexity=30, n_components=2, init='pca', n_iter=5000) # TSNE降维，降到2 low_dim_embs = tsne.fit_transform(test_data) plt.title('tsne result') plt.scatter(low_dim_embs[:,0], low_dim_embs[:,1], marker='o', c=test_labels) plt.savefig("tsne.png")
参考：
https://blog.csdn.net/weixin_39373480/article/details/88742200
【深入浅出图神经网络】