基于LSTM网络的股价预测研究  

Research on Stock Price Prediction Based on LSTM Network

作　　者：黄玉成 方伟伟[1] Huang Yucheng;Fang Weiwei(Department of Information Engineering,Nanyang Institute of Technology,Nanyang 473000)

机构地区：[1]南阳理工学院信息工程学院,南阳473000

出　　处：《现代计算机》2021年第34期51-55,60,共6页Modern Computer

基　　金：河南省科技计划项目(182102210459)。

摘　　要：随着科技水平的提高和和经济市场的展开,传统的股价分析方法已经不能精准的预测未来股票价格趋势,影响投资者们对股价的走势判别。长短期记忆网络(long short-term memory,LSTM)作为一种基于深度学习中的循环神经网络(rerrent neural network,RNN),可以处理RNN带来的梯度问题,特别适合处理具有长期依赖关系的数据。本文应用LSTM网络,通过对以往股票价格的关系建模来预测未来的股价。实验对代号002的上证股票价格进行预测,实验结果表明,选用LSTM模型来进行股票价格的预测,其预测的结果误差小,精准度极高。With the improvement of science and technology and the development of the economic market,the traditional stock price analysis method cannot accurately predict the future stock price trend,which affects investors'judgment of the stock price trend.Long short-term memory(LSTM)network is a recurrent neural network(RNN)based on deep learning,which can deal with the gra⁃dient problem caused by RNN,and is especially suitable for dealing with data with long-term dependence.In this paper,LSTM net⁃work is used to predict the future stock price by modeling the relationship between the past stock price.The dataset uses the stock data of Shanghai stock exchange code 002.The experimental results show that the LSTM model has smaller error and higher accuracy in stock price prediction.

关 键 词：LSTM 时间序列长度 股票价格预测 

分 类 号：F832.51[经济管理—金融学] F224