摘要
本文根据股指、股价等数据的时序特征将人工神经网络(ANN)与深度学习中的循环神经网络(RNN)引入股指预测,基于BP神经网络模型与长短期记忆(LSTM)神经网络模型构建了BP-LSTM模型.基于上证指数,本文进行了进行数值实验.结果表明BP-LSTM预测模型的准确率相比传统机器学习模型有明显提升,与普通LSTM模型相比也有较大提升.
In this paper,according to the time series characteristics of financial data such as stock index and stock price,and introducing the Artificial Neural Network(ANN)and Recurrent Neural Network(RNN)in deep learning to stock index prediction,we build the BP-LSTM model based on Back Propagation(BP)neural network model and Long Short-Term Memory(LSTM)neural network model.Numerical analysis shows that the accuracy of BP-LSTM prediction model is significantly higher than the traditional machine learning models and the LSTM model.
作者
孙存浩
胡兵
邹雨轩
SUN Cun-Hao;HU Bing;ZOU Yu-Xuan(School of Mathematics,Sichuan University,Chengdu 610064,China)
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2020年第1期27-31,共5页
Journal of Sichuan University(Natural Science Edition)
基金
国家自然科学基金(11401407)
关键词
BP神经网络
长短期记忆神经网络
上证指数趋势预测
Back propagation neural network
Long short-term memory neural network
Shanghai composite index trend forecast