两种基于深度网络的股票价格预测方法研究
Research on Two Stock Price Forecasting Methods Based on Deep Network
孙震宇1
作者信息
- 1. 云南师范大学 数学学院,云南 昆明 650500;云南省现代分析数学及其应用重点实验室,云南 昆明 650500
- 折叠
摘要
股票是一种重要的投资渠道,如何更准确地预测股票价格是一个热门的研究课题.由于股票数据的非线性、非平稳以及前后相关等复杂特点,传统的股票价格预测方法已经到达性能瓶颈.随着深度学习方法的兴起,LSTM和GRU等深度神经网络预测模型受到了极大的关注.基于厦门港务股票和上证指数的历史交易数据,利用了LSTM和GRU两种模型对收盘价进行预测研究,通过5个指标MAE、MSE、RMSE、MAPE和R2 给出了模型评价.
Abstract
Stock is an important investment channel,how to forecast stock price more accurately is a hot research topic.Due to the complex characteristics of stock data,such as non-linearity,non-stationarity and before and after correlation,traditional stock price forecasting methods have reached the performance bottleneck.With the rise of Deep Learning methods,deep neural network forecast models such as LSTM and GRU have received great attention.Based on the historical trading data of Xiamen Port Stock and Shanghai Stock Index,LSTM and GRU models are used to forecast the closing price.The model evaluation is given by 5 indexes of MAE,MSE,RMSE,MAPE and R2.
关键词
股票价格预测/LSTM模型/GRU模型Key words
stock price prediction/LSTM model/GRU model引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据