基于深度学习的LSTM模型在股票价格预测中的应用

Application of LSTM Model Based on Deep Learning in Stock Price Prediction

摘要: 股票价格波动与投资者利益、宏观经济稳定密切相关，精准预测股票价格备受关注。传统的预测方法，如循环神经网络(RNN)模型在预测非线性、非正态和时间相关性的股票数据时存在长期记忆能力较差、效率低下等问题。长短期记忆网络(LSTM)模型作为改进的循环神经网络，能更好地处理长时间跨度上的相关性，提高预测的准确性和稳定性。本文以上证综指为研究对象，将LSTM模型与RNN模型进行对比实验，评估分析模型性能，结果表明使用LSTM模型对股价的预测误差小，效果佳，在股票价格预测领域具有一定的应用价值。

Abstract: The fluctuations of stock prices are closely related to the interests of investors and the stability of the macro-economy. Precise prediction of stock prices has attracted much attention. Traditional prediction methods, such as the Recurrent Neural Network (RNN) model, have problems like poor long-term memory ability and low efficiency when predicting stock data that is nonlinear, non-normal, and has time correlations. As an improved recurrent neural network, the Long Short-Term Memory (LSTM) model can better handle the correlations over a long time span, improving the accuracy and stability of predictions. This paper takes the Shanghai Composite Index as the research object, conducts a comparative experiment between the LSTM model and the RNN model, and evaluates and analyzes the performance of the models. The results show that the predictions of stock prices using the LSTM model have small errors and good effects, indicating that the LSTM model has certain application value in the field of stock price prediction.