pytorch RNN股票预测 示例

PyTorch是一个流行的深度学习库，常用于构建循环神经网络（RNNs），特别是用于序列数据如时间序列预测。股票价格预测是其中一个应用场景。下面是一个简单的例子，展示如何使用PyTorch构建一个基本的LSTM模型来进行股票价格预测：

import torch
from torch import nn

# 假设我们有训练集X_train和对应的收盘价y_train
class StockPricePredictor(nn.Module):
    def __init__(self, input_size, hidden_size, num_layers, output_size):
        super(StockPricePredictor, self).__init__()
        self.rnn = nn.LSTM(input_size, hidden_size, num_layers, batch_first=True)
        self.fc = nn.Linear(hidden_size, output_size)

    def forward(self, x):
        # x shape (batch_size, sequence_length, input_size)
        out, _ = self.rnn(x)  # 只取最后一个时间步的隐藏状态
        out = self.fc(out[:, -1, :])  # 取最后一层的输出
        return out

model = StockPricePredictor(input_size, hidden_size=64, num_layers=1, output_size=1)
optimizer = torch.optim.Adam(model.parameters())
criterion = nn.MSELoss()  # 使用均方误差作为损失函数

for epoch in range(num_epochs):  
    # 预测并计算损失
    predictions = model(X_train)
    loss = criterion(predictions, y_train)

    # 反向传播和优化
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()

# 测试阶段
model.eval()
with torch.no_grad():
    predicted_prices = model(X_test).detach().numpy()