计算机应用与软件2018，Vol.35Issue(4)：44-48,108,6.DOI:10.3969/j.issn.1000-386x.2018.04.008
基于正则化LSTM模型的股票指数预测
STOCK INDEX FORECAST BASED ON REGULARIZED LSTM MODEL
摘要
Abstract
Aiming at the problem of financial time series forecasting,a new model of Long Short-Term Memory neural network (LSTM) was proposed.The LSTM model could dig out the inherent laws in time series through its unique element structure.The regularization method was used to modify the objective function of the LSTM model to optimize the network structure,so as to select the elasticized regularized LSTM model with generalized ability.The model was applied to the Dow Jones index forecast,and the experimental results showed that the proposed method had the lowest root mean square error and the highest prediction fitting degree.关键词
LSTM模型/正则化方法/股票指数/预测Key words
LSTM model/Regularized method/Stock index/Prediction分类
信息技术与安全科学引用本文复制引用
任君,王建华,王传美,王建祥..基于正则化LSTM模型的股票指数预测[J].计算机应用与软件,2018,35(4):44-48,108,6.基金项目
教育部人文社科青年基金项目(14YJCZH143) （14YJCZH143）
中央高校基本科研业务费专项(WUT:2016IA005). （WUT:2016IA005）
