本文为美国密苏里大学（作者：Qiyuan Gao）的硕士论文，共55页。

在本文中，我们研究了利用长短期记忆（LSTM）的递归神经网络（RNN）进行股市预测的问题。本研究旨在探讨LSTM在股市预测中的可行性及表现，我们通过测试不同的配置来优化LSTM模型，即隐层中的神经元数量和序列中的样本数量。我们不使用每日股票价格数据，而是从IQFEED数据库收集每小时股票数据，以便用相对较低的噪声样本来训练我们的模型。基于LSTM模型的预测结果，我们建立了一个由五个不同行业的6只美国市场股票组成的股票数据库。这六种股票的平均测试准确度为54.83%，最高的准确度为59.5%，最低的准确度为49.75%。然后，我们开发了一个交易模拟器，通过在400小时内对投资组合进行投资来评估我们的模型性能，该模型获得的总利润为413233.33美元，初始投资为6000000美元。

In this research, we study the problem ofstock market forecasting using Recurrent Neural Network(RNN) with LongShort-Term Memory (LSTM). The purpose of this research is to examine thefeasibility and performance of LSTM in stock market forecasting. We optimizethe LSTM model by testing different configurations, i.e., the number of neuronsin hidden layers and number of samples in sequence. Instead of using dailystock price data, we collect hourly stock data from the IQFEED database inorder to train our model with relatively low noise samples. Nevertheless, basedon the prediction results of LSTM model, we build up a stock database with sixU.S market stocks from five different industries. The average test accuracy ofthese six stocks is 54.83%, where the highest accuracy is at 59.5% while thelowest is at 49.75%. We then develop a trade simulator to evaluate theperformance of our model by investing the portfolio within a period of 400hours, the total profit gained by the model is $413,233.33 with $6,000,000initial investment.

1 引言

2 文献回顾

3 算法设计

4 具体实现

5 实验与结果

6 结论与未来工作展望

更多精彩文章请关注公众号：