Multi-Scale Stock Prediction Based on Deep Transfer Learning

doi:10.3778/j.issn.1002-8331.2112-0501

Abstract

Abstract: The stock market is not only an important financing channel for listed companies, but also an important investment market. Stock prediction has always attracted people’s attention. In order to make full use of the information from different stock data and improve the prediction performance of stocks, a multi-scale stock price prediction model TL-EMD-LSTM-MA（TELM） is proposed. TELM decomposes the closing price into multiple time-scale components through empirical mode decomposition. Different time-scale components have different oscillation frequencies, reflecting different periodic information. Different components are predicted by different method according to oscillation frequencies. High-frequency component are predicted by the stacked LSTM which is trained by deep transfer learning method. Low-frequency component are predicted by the moving average method. The predicted closing price is calculated by summing the predicted values of all components. The stacked LSTM trained by deep transfer learning contains information from different stocks and has more industry or market knowledge, which can effectively reduce the predictive error. Using the moving average method to predict the low-frequency component can more effectively capture the general trend of stocks. 500 stocks in China’s A-share market, Shanghai stock index, SZSE component index, and other stock indexes are predicted. The results show that the error of TELM is the lowest and the goodness of fit is the highest compared with other models. By joining in the stock predicted closing price of TELM, the stock trading process is simulated. The results show that TELM has low investment risk and high return.

Key words: stock prediction, deep transfer learning, empirical mode decomposition, long short-term memory, moving average

摘要： 股票市场不仅是上市公司的重要融资渠道，也是重要的投资市场，股票预测一直受到人们的关注。为了充分利用来自不同股票价格的信息，提高股票的预测效果，提出一种多尺度股票价格预测模型TL-EMD-LSTM-MA（TELM）。TELM模型通过经验模态分解将收盘价分解为多个时间尺度分量，不同时间尺度分量震荡频率不同，反映了不同的周期性信息；根据分量的震荡频率选择不同方法进行预测，高频分量利用深度迁移学习的方法训练堆叠LSTM，低频分量利用移动平均法进行预测；将所有分量的预测值相加作为收盘价的最终预测输出。通过深度迁移学习训练的堆叠LSTM，包含来自不同股票的信息，具备更多行业或市场的知识，能有效降低预测误差。利用移动平均法预测低频分量，更有效捕获股票的总体趋势。对中国A股市场内500支股票以及上证指数、深证成指等指数进行预测，结果表明，与其他模型相比，TELM预测误差最低，拟合优度最高。根据TELM预测的股票收盘价模拟股票交易过程，结果表明TELM投资风险低、收益高。

关键词: 股票预测, 深度迁移学习, 经验模态分解, 长短期记忆网络, 移动平均方法

CHENG Mengfei, GAO Shuping. Multi-Scale Stock Prediction Based on Deep Transfer Learning[J]. Computer Engineering and Applications, 2022, 58(12): 249-259.

程孟菲, 高淑萍. 基于深度迁移学习的多尺度股票预测[J]. 计算机工程与应用, 2022, 58(12): 249-259.

References

[1] SONG F M，LI H Y.Volatility decomposition of an empirical study on China stock market[J].Collected Essays on Finance and Economics，2004（4）：23-28.
[2] SHARMA A，BHURIYA D，SINGH U.Survey of stock market prediction using machine learning approach[C]//2017 International Conference of Electronics，Communication and Aerospace Technology（ICECA），2017：506-509.
[3] 娄磊，刘璐，刘先俊，等.基于小波去噪的ARIMA-LSTM混合模型及对股票价格指数的预测[J].长春理工大学学报（自然科学版），2021，44（2）：119-123.
LOU L，LIU L，LIU X J，et al.ARIMA-LSTM hybrid model based on wavelet denoising and prediction of stock price index[J].Journal of Changchun University of Science and Technology（Natural Science Edition），2021，44（2）：119-123.
[4] 贺毅岳，高妮，王峰虎，等.EMD分解下基于SVR的股票价格集成预测[J].西北大学学报（自然科学版），2019，49（3）：329-336.
HE Y Y，GAO N，WANG F H，et al.Research on integrated forecasting of stock price based on EMD and support vector regression[J].Journal of Northwest University（Natural Science Edition），2019，49（3）：329-336.
[5] ZHANG X D，ZHAI R，GAO W.Analysis and economic prediction of stock index based on index tracking and ARIMA model[J].Journal of Physics：Conference Series，2021，1903（1）.
[6] MALLQUI D C A，FERNANDES R A S.Predicting the direction，maximum，minimum and closing prices of daily Bitcoin exchange rate using machine learning techniques[J].Applied Soft Computing，2019，75：596-606.
[7] RIBEIRO M H D M，DOS SANTOS COELHO L.Ensemble approach based on bagging，boosting and stacking for short-term prediction in agribusiness time series[J].Applied Soft Computing，2020，86：105837.
[8] DASH R，DASH P K.A hybrid stock trading framework integrating technical analysis with machine learning techniques[J].The Journal of Finance and Data Science，2016，2（1）：42-57.
[9] YAN X，CAI Q，ZHANG S，et al.Exploring machine learning in stock prediction using LSTM，binary tree，and linear regression algorithms[J].International Core Journal of Engineering，2021，7（3）：373-377.
[10] ALTHELAYA K A，EL-ALFY E，MOHAMMED S.Stock market forecast using multivariate analysis with bidirectional and stacked（LSTM，GRU）[C]//2018 21st Saudi Computer Society National Computer Conference（NCC），2018：1-7.
[11] ZHAO Z，RAO R，TU S，et al.Time-weighted lstm model with redefined labeling for stock trend prediction[C]//2017 IEEE 29th International Conference on Tools with Artificial Intelligence（ICTAI），2017：1210-1217.
[12] MOOTHA S，SRIDHAR S，SEETHARAMAN R，et al.Stock price prediction using bi-directional LSTM based sequence to sequence modeling and multitask learning[C]//2020 11th IEEE Annual Ubiquitous Computing，Electronics & Mobile Communication Conference（UEMCON）.New York City，USA：IEEE，2020：78-86.
[13] ALTHELAYA K A，EL-ALFY E M，MOHAMMED S.Evaluation of bidirectional LSTM for short-and long-term stock market prediction[C]//2018 9th International Conference on Information and Communication Systems（ICICS），2018：151-156.
[14] QIAN F，CHEN X.Stock prediction based on LSTM under different stability[C]//2019 IEEE 4th International Conference on Cloud Computing and Big Data Analysis（ICCCBDA），2019：483-486.
[15] LI B，LI L F.Stock prediction based on adaptive gradient descent deep learning[M].[S.l.]：Springer International Publishing，2021：51-62.
[16] NGUYEN T T，YOON S.A novel approach to short-term stock price movement prediction using transfer learning[J].Applied Sciences，2019，9（22）：4745.
[17] 张倩玉，严冬梅，韩佳彤.结合深度学习和分解算法的股票价格预测研究[J].计算机工程与应用，2021，57（5）：56-64.
ZHANG Q Y，YAN D M，HAN J T.Research on stock price prediction combined with deep learning and decomposition algorithm[J].Computer Engineering and Applications，2021，57（5）：56-64.
[18] LIU F，LI X，WANG L.Exploring cluster stocks based on deep learning for stock prediction[C]//2019 12th International Symposium on Computational Intelligence and Design（ISCID），2019：107-110.
[19] HUANG N E，SHEN Z，LONG S R，et al.The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J].Proceedings Mathematical Physical & Engineering Sciences，1998，454（1971）：903-995.
[20] KUMAR P，HATI A S.Transfer learning-based deep CNN model for multiple faults detection in SCIM[J].Neural Computing and Applications，2021，33：15851-15862.
[21] PAN S J，YANG Q.A survey on transfer learning[J].IEEE Transactions on Knowledge and Data Engineering，2010，22（10）：1345-1359.
[22] HASSAN D S U，MUZAFFER ?，BURAK ? A，et al.A transfer-learning approach for accelerated MRI using deep neural networks[J].Magnetic Resonance in Medicine，2020，84（2）：663-685.
[23] WANG W，WANG Z，ZHOU Z，et al.Anomaly detection of industrial control systems based on transfer learning[J].Tsinghua Science and Technology，2021，26（6）：821-832.
[24] TAN C，SUN F，KONG T，et al.A survey on deep transfer learning[C]//Artificial Neural Networks and Machine Learning（ICANN 2018）.Cham：Springer International Publishing，2018：270-279.
[25] HOCHREITER S，SCHMIDHUBER J.Long shart-term memory[J].Neural Computation，1997，9（8）：1735-1780.
[26] CHIARELLA C，HE X Z，HOMMES C.A dynamic analysis of moving average rules[J].Journal of Economic Dynamics and Control，2006，30（9/10）：1729-1753.
[27] 曲健，陈红岩，刘文贞，等.基于改进网格搜索法的支持向量机在气体定量分析中的应用[J].传感技术学报，2015，28（5）：774-778.
QU J，CHEN H Y，LIU W Z，et al.Application of support vector machine based on improved grid search in quantitative analysis of gas[J].Chinese Journal of Sensors and Actuators，2015，28（5）：774-778.