Application of Deep Learning in Stock Investment

DOI: https://doi.org/10.62517/jike.202604226

Author(s)

Shengze Wu

Affiliation(s)

College of Marine Information Engineering, Jimei University, Xiamen, Fujian, China

Abstract

Deep learning continues to reshape financial modeling, showing significant promise for stock investment decisions. This paper evaluates the practical viability of these techniques by utilizing Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN) to model price predictions and optimize portfolios. Drawing on CSI 300 index constituents (2018–2023), we engineered a multidimensional dataset capturing technicals, fundamentals, and market sentiment. Tests show the LSTM model reached 73.5% accuracy in short-term forecasts, notably beating traditional ARIMA and Support Vector Machine baselines. For portfolio construction, coupling deep reinforcement learning with modern portfolio theory yielded an annualized return of 14.2% and a Sharpe ratio of 1.85, outperforming the benchmark by 5.8%. During market turbulence, these deep learning frameworks displayed superior stability and adaptability. We also address inherent overfitting and black-box limitations, noting that regularization and interpretability tools are essential for future refinement. Ultimately, this work offers concrete guidance for deploying AI in real-world financial contexts.

Keywords

Deep Learning; Stock Investment; Price Prediction; Portfolio Optimization; Financial Technology

References

[1] Krauss, C., Do, X. A., & Huck, N. (2017). Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500. European Journal of Operational Research, (02), 689-702. [2] Cont, R. (2001). Empirical properties of asset returns: Stylized facts and statistical issues. Quantitative Finance, (02), 223-236. [3] Sezer, O. B., Gudelek, M. U., & Ozbayoglu, A. M. (2020). Financial time series forecasting with deep learning: A systematic literature review: 2005-2019. Applied Soft Computing, (90), 106181. [4] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, (08), 1735-1780. [5] Lu, W., Li, J., Li, Y., Sun, A., & Wang, J. (2020). A CNN-LSTM-based model to forecast stock prices. Complexity, (01), 6622927. [6] Zuo, C., Wang, Y., & Li, Z. (2022). Application of deep learning in stock investment. Journal of Zhengzhou University (Philosophy and Social Sciences Edition), (02), 57-63. [7] Song, F., Tian, C., & Feng, C. (2025). Research on quantitative stock investment based on deep reinforcement learning. Journal of Nanjing Normal University (Natural Science Edition), (01), 91-98. [8] Zuo, X. (2023). Research on stock portfolio optimization based on investor sentiment and deep reinforcement learning. China Management Informationization, (03), 132-135. [9] Ma, T., Jiang, F., & Tang, G. (2022). Deep learning and factor investment in China's stock market: A generative adversarial network approach. China Economic Quarterly, (03), 97-120. [10] Pei, M., & Zhang, L. (2020). Research on stock price trend prediction based on investor sentiment and deep learning. Financial Development, (01), 82-101. [11] Dong, J. (2022). Current status and development trends of deep learning applications in quantitative investment at home and abroad. Modern Business, (01), 122-124. [12] Xie, J., Huang, B., Chen, S., Peng, Q., & Zheng, W. (2020). Application of deep learning technology in investor suitability management. Financial Perspective, (06), 35-45. [13] Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, (02), 654-669. [14] Deng, Y., Bao, F., Kong, Y., Ren, Z., & Dai, Q. (2017). Deep direct reinforcement learning for financial signal representation and trading. IEEE Transactions on Neural Networks and Learning Systems, (03), 653-664.