帮我写一段GNN预测用户违约风险的python代码

好的，以下是一个简单的GNN预测用户违约风险的Python代码：

import torch
import torch.nn.functional as F
from torch_geometric.nn import GCNConv

class GCN(torch.nn.Module):
    def __init__(self, num_features, hidden_channels, num_classes):
        super(GCN, self).__init__()
        self.conv1 = GCNConv(num_features, hidden_channels)
        self.conv2 = GCNConv(hidden_channels, hidden_channels)
        self.conv3 = GCNConv(hidden_channels, hidden_channels)
        self.lin = torch.nn.Linear(hidden_channels, num_classes)

    def forward(self, x, edge_index):
        x = F.relu(self.conv1(x, edge_index))
        x = F.relu(self.conv2(x, edge_index))
        x = F.relu(self.conv3(x, edge_index))
        x = self.lin(x)
        return x

# 加载数据
x = ... # 特征矩阵
edge_index = ... # 边索引
y = ... # 标签

# 设置超参数
num_features = x.size(-1)
hidden_channels = 128
num_classes = 2
lr = 0.01
epochs = 100

# 初始化模型和优化器
model = GCN(num_features, hidden_channels, num_classes)
optimizer = torch.optim.Adam(model.parameters(), lr=lr)

# 训练模型
model.train()
for epoch in range(epochs):
    optimizer.zero_grad()
    out = model(x, edge_index)
    loss = F.cross_entropy(out, y)
    loss.backward()
    optimizer.step()

# 测试模型
model.eval()
out = model(x, edge_index)
pred = out.argmax(dim=1)
accuracy = (pred == y).sum().item() / y.size(0)
print('Accuracy: {:.4f}'.format(accuracy))

这段代码定义了一个3层的GCN模型，用于预测用户违约风险。输入数据包括特征矩阵x和边索引edge_index，标签y表示用户是否违约。模型的输出是一个2维张量，第一维表示未违约的概率，第二维表示违约的概率。训练过程中使用了交叉熵损失和Adam优化器，测试时计算了模型的准确率。