基于离散化和LSTM神经网络的股票指数趋势预测研究

Research on Stock Index Trend Prediction Based on Discretization and LSTM

股票市场预测一直是一项受到各个领域研究者关注且极具挑战性的任务。针对当前使用股票技术指标预测股市精度不高的问题,提出将连续型数值的股票技术指标特征离散化为一系列0、1特征,同时加入宏观经济指标的方法应用于股票指数预测。实验以沪深300成分指数为源数据,对沪深300成分指数的涨、跌进行预测,实验结果显示,离散化技术以及宏观经济指标均对股指趋势预测的精度有所提升。

Predicting stock market has attracted many researchers in multiple fields and challenging task.Aiming at the problem that current stock technical indicators is used to predict the accuracy of the stock market not well,proposes to discretize the continuous numerical stock index characteristics into a series of 0 and 1 features,and addition of macroeconomic variables to the stock index forecast.The experiment uses the CSI 300 market index as the source data to predict the rise and fall of the stock indices.The experimental results show that the discreti⁃zation technology and macroeconomic variables have improved the accuracy of the stock index trend forecast.