基于ARIAM-GARCH深度学习的股价预测与决策  

Stock Price Prediction and Decision Based on ARIAM-GARCH Deep Learning

作　　者：刘祺 施三支[1] 娄磊 刘璐[1,3] LIU Qi;SHI Sanzhi;LOU Lei;LIU Lu(School of Mathematics and Statistics,Changchun University of Science and Technology,Changchun 130022;Northeast Securities Co,Ltd.Changchun Ecological Street Business Department,Changchun 130022;Chuncheng Subdistricts of China,Lvyuan District,Changchun City,Changchun 130022)

机构地区：[1]长春理工大学数学与统计学院,长春130022 [2]东北证券股份有限公司长春生态大街营业部,长春130022 [3]长春市绿园区春城街道办事处,长春130022

出　　处：《长春理工大学学报（自然科学版）》2024年第1期119-130,共12页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金(11601039);吉林省教育厅项目(JJKH20210809KJ)。

摘　　要：对于绝大多数股民而言,大盘股具有投入成本高、损失大、资金流转慢等特性,为使得股民减少损失,提高收益,通过针对于中小企业股,采用两阶段的方法对股价进行了预测,给出了一种最优投资组合决策方法。第一阶段先用ARIMA-GARCH模型进行建模,提取每日收益率的波动率,再将波动率加入到原有数据集中,构建CNN-BiLSTM-AT深度神经网络模型进行预测,最后使用XGBoost算法对预测值进行修正。第二阶段在给定投资者的期望收益率的条件下,用Bayes方法进行投资组合,获取最优投资决策。实证研究表明,此方法对于选取的10只中小企业股有着较好的研究结果。For the vast majority of shareholders,large-cap stocks often represent characteristics such as high investment costs,large losses,and slow capital flow.In order to reduce losses and improve returns for shareholders,a two-stage method was used to predict stock prices for small and medium-sized enterprise stocks,and an optimal investment portfolio decision-making method was proposed.In the first stage,the ARIMA-GARCH model is used to model,extract the volatility of daily returns,and then add the volatility to the original dataset to construct a CNN-BiLSTM-AT deep neural network model for prediction.Finally,the XGBoost algorithm is used to correct the predicted values.In the second stage,given the expected return rate of investors,the Bayes method is used to make the investment portfolio and obtain the optimal investment decision.Empirical research shows that this method has good research results for the selected 10 small and medium-sized enterprise stocks.

关 键 词：股价预测 ARIMA-GARCH模型 CNN-BiLSTM-AT XGBoost算法 BAYES方法 

分 类 号：O482.31[理学—固体物理]