Stock trend prediction method coupled with multilevel indicators

Abstract

To systematically incorporate multiple influencing factors, the coupled-state frequency memory(Co-SFM)network is proposed. This model integrates Copula estimation with neural networks, fusing multilevel data information, which is then fed into downstream learning modules. Co-SFM employs an upstream fusion module to incorporate multilevel data, thereby constructing a macro-plate-micro data structure. This configuration helps identify and integrate characteristics from different data levels, facilitating a deeper understanding of the internal links within the financial system. In the downstream model, Co-SFM uses a state-frequency memory network to mine hidden frequency information within stock prices, and the multifrequency patterns of sequential data are modeled. Empirical results show that Co-SFM’s prediction accuracy for stock price trends is significantly better than that of other models. This is especially evident in multistep medium and long-term trend predictions, where integrating multilevel data results in notably improved accuracy.

Keywords

stock trend prediction / multilevel indicators / Copula / state-frequency memory network

References

[1] Malkiel B G. A random walk down wall street: Including a life-cycle guide to personal investing[M]. New York: Ww Norton & Company, 1990: 5-15.

[2] Ariyo A A, Adewumi A O, Ayo C K. Stock price prediction using the ARIMA model[C]// 2014 UKSim-AMSS 16th International Conference on Computer Modelling and Simulation. Cambridge, UK, 2014: 106-112. DOI: 10.1109/UKSim.2014.67.

[3] Russel E, Kesumah F S D, Rialdi A, et al. Dynamic modeling and forecasting stock price data by applying AR-GARCH model[J]. TEST Engineering and Management, 2020, 82: 6829-6842.

[4] Li S B, Kong X K, Li Q T, et al. Short-term traffic flow prediction with PSR-XGBoost considering chaotic characteristics[J]. Journal of Southeast University(English Edition), 2022, 38(1): 92-96. DOI: 10.3969/j.issn.1003-7985.2022.01.014.

[5] Wang Y H, He J Z, Zhang M Z, et al. Concrete crack identification in complex environments based on SSD and pruning neural network[J]. Journal of Southeast University(English Edition), 2023, 39(4): 393-399. DOI: 10.3969/j.issn.1003-7985.2023.04.008.

[6] Wang S, Zhang L J, Yin G J. Defect identification method for steel surfaces based on improved YOLOv5[J]. Journal of Southeast University(English Edition), 2024, 40(1): 49-57. DOI: 10.3969/j.issn.1003-7985.2024.01.006.

[7] Leung C K S, MacKinnon R K, Wang Y. A machine learning approach for stock price prediction[C]// Proceedings of the 18th International Database Engineering & Applications Symposium. Porto, Portugal, 2014: 274-277. DOI: 10.1145/2628194.2628211.

[8] Vijh M, Chandola D, Tikkiwal V A, et al. Stock closing price prediction using machine learning techniques[J]. Procedia Computer Science, 2020, 167: 599-606. DOI: 10.1016/j.procs.2020.03.326.

[9] Kim K J, Ahn H. Simultaneous optimization of artificial neural networks for financial forecasting[J]. Applied Intelligence, 2012, 36(4): 887-898. DOI: 10.1007/s10489-011-0303-2.

[10] Adebiyi A A, Ayo C K, Adebiyi M O, et al. Stock price prediction using neural network with hybridized market indicators[J]. Journal of Emerging Trends in Computing and Information Sciences, 2012, 3(1): 1-9.

[11] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.

[12] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30: 1-11. DOI: 10.48550/arXiv.1706.03762.

[13] Zhou H Y, Zhang S H, Peng J Q, et al. Informer: Beyond efficient transformer for long sequence time-series forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(12): 11106-11115. DOI: 10.1609/aaai.v35i12.17325.

[14] Niu Z Y, Zhong G Q, Yu H. A review on the attention mechanism of deep learning[J]. Neurocomputing, 2021, 452: 48-62. DOI: 10.1016/j.neucom.2021.03.091.

[15] He Z R, Shen Q F, Wu J X, et al. Transformer encoder-based multilevel representations with fusion feature input for speech emotion recognition[J]. Journal of Southeast University(English Edition), 2023, 39(1): 68-73. DOI: 10.3969/j.issn.1003-7985.2023.01.008.

[16] Niu Y L, Lu Z W, Wen J R, et al. Multi-modal multi-scale deep learning for large-scale image annotation[J]. IEEE Transactions on Image Processing: A Publication of the IEEE Signal Processing Society, 2019, 28(4): 1720-1731. DOI: 10.1109/TIP.2018.2881928.

[17] Liu M, Zhou W, Xu Z S. Hesitant fuzzy long short-term memory network and its application in the intelligent building selection[J]. IEEE Transactions on Fuzzy Systems, 2024, 32(5): 2590-2602. DOI: 10.1109/TFUZZ.2024.3355000.

[18] Li H, Zhao W, Zhang Y X, et al. Remaining useful life prediction using multi-scale deep convolutional neural network[J]. Applied Soft Computing, 2020, 89: 106113. DOI: 10.1016/j.asoc.2020.106113.

[19] Liu G, Mao Y, Sun Q, et al. Multi-scale two-way deep neural network for stock trend prediction[C]// Twenty-Ninth International Joint Conference on Artificial Intelligence and Seventeenth Pacific Rim International Conference on Artificial Intelligence(IJCAI). Stockholm, Sweden, 2020: 4555-4561. DOI: 10.24963/ijcai.2020/621.

[20] Chen Y X, Lin W W, Wang J Z. A dual-attention-based stock price trend prediction model with dual features[J]. IEEE Access, 2019, 7: 148047-148058. DOI: 10.1109/ACCESS.2019.2946223.

[21] Hou M, Xu C, Liu Y, et al. Stock trend prediction with multi-granularity data: A contrastive learning approach with adaptive fusion[C]// Proceedings of the 30th ACM International Conference on Information & Knowledge Management. Queensland, Australia, 2021: 700-709. DOI: 10.1145/3459637.3482483.

[22] Nourani V, Sharghi E, Behfar N, et al. Multi-step-ahead solar irradiance modeling employing multi-frequency deep learning models and climatic data[J]. Applied Energy, 2022, 315: 119069. DOI: 10.1016/j.apenergy.2022.119069.

[23] Wang G F, Cao L B, Zhao H K, et al. Coupling macro-sector-micro financial indicators for learning stock representations with less uncertainty[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(5): 4418-4426. DOI: 10.1609/aaai.v35i5.16568.

[24] Zhang L H, Aggarwal C, Qi G J. Stock price prediction via discovering multi-frequency trading patterns[C]// Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Halifax, NS, Canada, 2017: 2141- 2149. DOI: 10.1145/3097983.3098117.

[25] Sotomayor L R, Cadenillas A. Explicit solutions of consumption-investment problems in financial markets with regime switching[J]. Mathematical Finance, 2009, 19(2): 251-279. DOI: 10.1111/j.1467-9965.2009.00366.x.

[26] Ferson W E, Schadt R W. Measuring fund strategy and performance in changing economic conditions[J]. The Journal of Finance, 1996, 51(2): 425-461. DOI: 10.1111/j.1540-6261.1996.tb02690.x.

[27] Sklar A. Random variables, joint distribution functions, and copulas[J]. Kybernetika, 1973, 9(6): 449-460.

[28] Krishnamoorthy A, Menon D. Matrix inversion using Cholesky decomposition[C]//2013 Signal Processing: Algorithms, Architectures, Arrangements, and Applications(SPA). Poznan, Poland, 2013: 70-72.

Funding

The National Natural Science Foundation of China(72173018)