Research on Stock Price Prediction Based on Autoregressive Model of Maximum Corentropy Criterion

Shenghan Gao, Mengyang Liu, Lina Wang, Xiaoyan Qiao, Feng Zhao

2022

Abstract

As a barometer of the financial market, the stock market is closely related to national economic development, corporate financing and investors' interests. However, there are many and complex factors affecting stock price volatility, which makes accurate prediction of stock price volatility still a challenging problem. In order to predict the stock price more accurately, the maximum correlation entropy autoregression model is proposed in this paper. Specifically, the maximum entropy criterion is used to replace the minimum mean square error criterion in the autoregressive model to eliminate the influence of singular values. Then a new clustering method is used to cluster the segmented stock price curves, and a regression model is built for each class, which reduces the influence of the order of the regression model on the prediction accuracy. In addition, the open set identification method is adopted in this paper to add boundary constraints to each curve after clustering, which is used to enhance the pertinence of the regression model and effectively improve the prediction accuracy. The experimental results show that the proposed method has high prediction accuracy.

Paper Citation

in Harvard Style

Gao S., Liu M., Wang L., Qiao X. and Zhao F. (2022). Research on Stock Price Prediction Based on Autoregressive Model of Maximum Corentropy Criterion. In Proceedings of the 4th International Conference on Economic Management and Model Engineering - Volume 1: ICEMME; ISBN 978-989-758-636-1, SciTePress, pages 67-75. DOI: 10.5220/0012023700003620

in Bibtex Style

@conference{icemme22,
author={Shenghan Gao and Mengyang Liu and Lina Wang and Xiaoyan Qiao and Feng Zhao},
title={Research on Stock Price Prediction Based on Autoregressive Model of Maximum Corentropy Criterion},
booktitle={Proceedings of the 4th International Conference on Economic Management and Model Engineering - Volume 1: ICEMME},
year={2022},
pages={67-75},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012023700003620},
isbn={978-989-758-636-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 4th International Conference on Economic Management and Model Engineering - Volume 1: ICEMME
TI - Research on Stock Price Prediction Based on Autoregressive Model of Maximum Corentropy Criterion
SN - 978-989-758-636-1
AU - Gao S.
AU - Liu M.
AU - Wang L.
AU - Qiao X.
AU - Zhao F.
PY - 2022
SP - 67
EP - 75
DO - 10.5220/0012023700003620
PB - SciTePress