%0 Journal Article
%T 基于LSTM和GRU的股价预测研究<br>A Study on Stock Price Prediction Based on LSTM and GRU
%A 王代颖
%J E-Commerce Letters
%P 3203-3210
%@ 2168-5851
%D 2024
%I Hans Publishing
%R 10.12677/ecl.2024.132394
%X 随着计算机水平和数据库技术的迅猛发展，以神经网络为基础的深度学习已经成为机器学习领域中最热门的研究方向，人工神经网络、卷积神经网络、BP神经网络等先后被广泛应用于各领域中，并取得了很好的效果。如今，越来越多的金融学者也将目光投向了神经网络，运用网络技术建立股票收盘价趋势预测模型，取得了显著效果。本文基于用python构建了门控循环单元模型(GRU)和长短期时间记忆(LSTM)模型对光大银行收盘价进行预测，并对两种算法进行比较。实证结果表明：GRU模型在MAE、RMSE、MAPE三个评价指标的预测精度均高于LSTM模型，测试集的拟合也是最好的，能够在一定程度上反应股票价格波动趋势，为投资者进行投资决策提供了一个简单实用的方法。<br />
With the rapid development of computer level and database technology, deep learning based on neural network has become the hottest research direction in the neighborhood of machine learning, and artificial neural network, convolutional neural network, BP neural network and so on have successively been widely used in various fields and achieved good results. Nowadays, more and more financial scholars also focus on neural networks, using network technology to establish stock closing price trend prediction model, and have achieved significant results. This paper is based on the gated recurrent unit model (GRU) and long short-term time memory (LSTM) model constructed in python to predict the closing price of Everbright Bank, and compare the two algorithms. The empirical results show that the prediction accuracy of the GRU model in the three evaluation indicators of MAE, RMSE and MAPE is higher than that of the LSTM model, and the fit of the test set is also the best, which is able to respond to the trend of the stock price fluctuations to a certain extent, and provides a simple and practical for investors to make investment decisions.
%K 门控循环单元模型，长短期时间记忆，趋势预测<br>Gated Cyclic Cell Model
%K Long and Short-Term Time Memory
%K Trend Prediction
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=88045