MC_SVR滚动模型对股票价格的预测研究

doi:10.3778/j.issn.1002-8331.2401-0105

摘要/Abstract

摘要： 为了解决单一的可尔可夫链（Markov chain，MC）模型在预测股价时出现预测状态不准确，使得预测股价与实际股价相对误差过大，从而导致模型预测效果较差等问题，引入支持向量回归（support vector regression，SVR）模型，并结合滚动预测的思想，形成MC_SVR滚动模型。通过海泰发展的股票价格数据构建MC_SVR滚动模型，采用网格搜索法确定高斯核函数，惩罚系数[C=]204.003?906，核函数参数[γ=]0.003?906和损失函数参数[ε=]0.1。实验结果表明，MC_SVR滚动模型有效提高了预测结果的精度，相比于SVR模型和LSTM模型，平均绝对百分比误差[δ]分别降低了0.16和0.01个百分点，均方根误差RMSE分别降低了0.000?7和0.001?6，决定系数R2分别提高了0.000?8和0.001?8，DA统计量比SVR模型降低了1.908?7，比LSTM模型提高了8.227?8，从整体上表面MC_SVR滚动模型具有不错的预测精度。在新增10只股票的预测研究中，MC_SVR滚动模型均具有可行性和有效性。

关键词: 马尔可夫链, 核函数, 支持向量回归, 股票收盘价, 股价预测

Abstract: In order to solve the problem that the prediction state of a single MC (Markov chain) model is inaccurate in predicting stock prices, which causes the large relative error between the predicted stock price and the actual one, and results in poor prediction effect of the model, this paper introduces the SVR (support vector regression) model. By combined with the idea of rolling prediction, this paper obtains the MC_SVR rolling model. The MC_SVR rolling model is constructed through the stock price data of Haitai Development, and the grid search method is used to determine the Gaussian kernel function, with the penalty coefficient C=204.003?906, the kernel function parameter [γ]=0.003?906, and the loss function parameter [ε]=0.1. The experimental results indicate that the MC_SVR rolling model effectively improves the accuracy of prediction results. Compared with the SVR model and the LSTM model, the average absolute percentage error [δ] is reduced by 0.16 and 0.01?percentage points respectively, and the root mean square error (RMSE) is reduced by 0.000?7 and 0.001?6 respectively. The coefficient of determination R2 is increased by 0.000?8 and 0.001?8 respectively, and the DA statistic is 1.908?7 lower than that of the SVR model, and 8.227?8 higher than that of the LSTM model. Overall, it appears that the MC_SVR rolling model has good prediction accuracy. In the prediction research of 10 new stocks, the MC_SVR rolling model is feasible and effective.

Key words: Markov chain, kernel function, support vector regression, stock closing price, prediction of stock price

陈梓海, 黄香香. MC_SVR滚动模型对股票价格的预测研究[J]. 计算机工程与应用, 2025, 61(10): 166-175.

CHEN Zihai, HUANG Xiangxiang. Prediction Research on Stock Price of MC_SVR Rolling Model[J]. Computer Engineering and Applications, 2025, 61(10): 166-175.

参考文献

[1] 徐浩然, 许波, 徐可文. 机器学习在股票预测中的应用综述[J]. 计算机工程与应用, 2020, 56(12): 19-24.
XU H R, XU B, XU K W. Analysis on application of machine learning in stock forecasting[J]. Computer Engineering and Applications, 2020, 56(12): 19-24.
[2] 姚宏亮, 张远涛, 王浩, 等. 基于理性指标的马尔可夫链股市态势预测方法[J]. 计算机工程与应用, 2017, 53(22): 227-234.
YAO H L, ZHANG Y T, WANG H, et al. Markov chain stock market trend prediction method based on rational indicator[J]. Computer Engineering and Applications, 2017, 53(22): 227-234.
[3] D’AMICO G, DE BLASIS R. A multivariate Markov chain stock model[J]. Scandinavian Actuarial Journal, 2020(4): 272-291.
[4] XIE W S, METAWA N. Financial stock market forecasting based on intelligent support vector machine regression model[J]. Journal of Intelligent & Fuzzy Systems, 2021. DOI:10. 3233/JIFS-189839.
[5] 邓佳丽, 赵凤群, 王小侠. MTICA-AEO-SVR股票价格预测模型[J]. 计算机工程与应用, 2022, 58(8): 257-263.
DENG J L, ZHAO F Q, WANG X X. MTICA-AEO-SVR model for stock price forecasting[J]. Computer Engineering and Applications, 2022, 58(8): 257-263.
[6] 张冰, 王传美, 贺素香. 改进的TSVR模型在股市高频数据上的预测[J]. 计算机工程与设计, 2019, 40(11): 3241-3246.
ZHANG B, WANG C M, HE S X. Prediction of improved TSVR model on high frequency stock market[J]. Computer Engineering and Design, 2019, 40(11): 3241-3246.
[7] WU J M, LI Z C, HERENCSAR N, et al. A graph-based CNN-LSTM stock price prediction algorithm with leading indicators[J]. Multimedia Systems, 2023, 29(3): 1751-1770.
[8] 史建楠, 邹俊忠, 张见, 等. 基于DMD-LSTM模型的股票价格时间序列预测研究[J]. 计算机应用研究, 2020, 37(3): 662-666.
SHI J N, ZOU J Z, ZHANG J, et al. Research of stock price prediction based on DMD-LSTM model[J]. Application Research of Computers, 2020, 37(3): 662-666.
[9] 赵红蕊, 薛雷. 基于LSTM-CNN-CBAM模型的股票预测研究[J]. 计算机工程与应用, 2021, 57(3): 203-207.
ZHAO H R, XUE L. Research on stock forecasting based on LSTM-CNN-CBAM model[J]. Computer Engineering and Applications, 2021, 57(3): 203-207.
[10] 杨赟, 张丽丽. 基于ISOA-SVR模型的短期网络舆情预测[J]. 计算机工程与设计, 2024, 45(1): 168-176.
YANG Y, ZHANG L L. Short-term network public opinion prediction based on improved seagull optimization algorithm and support vector regression model[J]. Computer Engineering and Design, 2024, 45(1): 168-176.
[11] YE T. Stock forecasting method based on wavelet analysis and ARIMA-SVR model[C]//Proceedings of the 2017 3rd International Conference on Information Management. Piscataway: IEEE, 2017: 102-106.
[12] LIHKI R, KEYLA A. Forecasting selected Colombian shares using a hybrid ARIMA-SVR model[J]. Mathematics, 2022, 10(13): 2181.
[13] 周璇, 杨建成. 基于支持向量回归机的空调逐时负荷滚动预测算法[J]. 中南大学学报(自然科学版), 2014, 45(3): 952-957.
ZHOU X, YANG J C. An algorithm for hourly load rolling forecasting of air conditioning system based on SVR[J]. Journal of Central South University (Science and Technology), 2014, 45(3): 952-957.
[14] 刘致谦. 基于马尔可夫链的股票价格波动预测[J]. 合作经济与科技, 2021(24): 64-67.
LIU Z Q. Stock price fluctuation prediction based on Markov chain[J]. Co-Operative Economy & Science, 2021(24): 64-67.
[15] 郭佳, 马朝斌, 张绍博, 等. 优化的马尔可夫链人工蜂群算法[J]. 计算机科学与探索, 2020, 14(6): 985-995.
GUO J, MA C B, ZHANG S B, et al. Optimized artificial bee colony algorithm with Markov chain[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(6): 985-995.
[16] 李航. 统计学习方法[M]. 2版. 北京: 清华大学出版社, 2019: 355-356.
LI H. Statistical learning method[M]. 2nd ed. Beijing: Tsing-hua University Press, 2019: 355-356.
[17] 张伟. 基于马尔可夫链的郑州地区降雨量预测研究[D]. 郑州: 华北水利水电大学, 2019.
ZHANG W. Rainfall prediction study in Zhengzhou area based on Markov chain[D]. Zhengzhou: North China University of Water Resources and Electric Power, 2019.
[18] 徐熠明, 赵寿为. 改良马尔可夫模型在节假日客流预测中的应用[J]. 软件导刊, 2021, 20(9): 174-179.
XU Y M, ZHAO S W. Application of improved Markov model in holiday passenger flow prediction[J]. Software Guide, 2021, 20(9): 174-179.
[19] 苗正伟, 徐利岗. 基于隶属度修正的加权马尔可夫链的降水预测[J]. 长江科学院院报, 2018, 35(1): 40-46.
MIAO Z W, XU L G. Prediction of annual precipitation by weighted Markov chain based on membership modification[J]. Journal of Yangtze River Scientific Research Institute, 2018, 35(1): 40-46.
[20] 刘家学, 白明皓, 郝磊. 基于ARIMA-SVR组合方法的航班滑出时间预测[J]. 中国科技论文, 2021, 16(6): 661-667.
LIU J X, BAI M H, HAO L. Prediction of flight taxi-out time based on ARIMA-SVR combination method[J]. China Sciencepaper, 2021, 16(6): 661-667.
[21] VAPNIK V N. An overview of statistical learning theory[J]. IEEE Transactions on Neural Networks, 1999, 10(5): 988-999.
[22] 陈梓海, 黄香香. 基于加权马尔可夫链的股票价格预测研究[J]. 东莞理工学院学报, 2023, 30(3): 1-8.
CHEN Z H, HUANG X X. Research on forecasting stock price based on weighted Markov chain[J]. Journal of Dongguan University of Technology, 2023, 30(3): 1-8.