Theoretical Framework and Empirical Analysis of Stock Price and Trend Prediction Using LSTM Models

Xiaoyu Yang

doi:10.54097/b71tn654

Authors

Xiaoyu Yang

DOI:

https://doi.org/10.54097/b71tn654

Keywords:

Long Short-Term Memory, Stock Price Prediction, Time Series, Deep Learning.

Abstract

Changes in stock market prices are nonlinear, high-noise, and they change with time, which makes it hard for typical time series models to accurately capture their complicated patterns. This paper develops a forecasting model for stock prices using Long Short-Term Memory (LSTM) networks, with a focus on the A-share market to thoroughly investigate the effectiveness of LSTM models in predicting stock prices and trends. The research employs the Kaggle dataset titled “S&P 500 Stock Data” covering the years 2013 to 2018, which includes key daily trading indicators like opening price and trading volume. When comparing the predictive performance of LSTM models to that of traditional machine learning models like random forests, it is clear that LSTM models are better at predicting short-term trends and prices than both traditional machine learning models and traditional time series models. This research offers novel analytical insights for forecasting financial markets and possesses practical implications for investor decision-making and risk management.

Downloads

Download data is not yet available.

References

[1] Gandhmal, D. P., Kumar, K. Systematic analysis and review of stock market prediction techniques. Computer Science Review, 34:100190 (2019).

[2] Di Persio, L., Honchar, O. Recurrent Neural Networks Approach to the Financial Forecast of Google Assets. International Journal of Mathematics and Computers in Simulation, 11:7-13 (2017).

[3] Liu, G. Q. Research on the application of the nonlinear GARCH model in predicting Chinese stock market activities. Statistical Research, 2000 (1):49-52 (2000).

[4] Shi, Z. MambaStock: Selective state space model for stock prediction. arXiv (2024). https://doi.org/10.48550/arXiv.2402.18959

[5] 5) Shi, T. Study of short-term load forecasting based on random forest regression algorithm. Master's thesis, Zhengzhou University (2017).

[6] Hu, D. G., Ma, Y., Pang, X. D., Wu, F., Niu, Z., Li, H., Ji, Y. Q., Feng, W. T. New energy power generation prediction based on the CNN-LSTM-Attention model and risk detection analysis of isolation forest algorithm. Journal of Image and Signal Processing, 14 (1):45-61 (2025). https://doi.org/10.12677 /jisp.2025.141005

[7] Hochreiter, S., Schmidhuber, J. Long Short-term Memory. Neural Computation, 8:1735-1780 (1997).

[8] Gers, F. A., Schmidhuber, J., Cummins, F. Learning to forget: Continual prediction with LSTM. In Artificial Neural Networks (pp. 850-855). IEE (1999). https://doi.org/10.1049/cp:19991117

[9] Jiang, S. Y. Stock price prediction based on LSTM model. Jiangsu Commercial Forum, 2025 (1):83-86 (2025). https://doi.org/10.13395/j.cnki.issn.1009-0061.2025.01.013

[10] Murphy, J. J. Technical analysis of the financial markets. Seismological Press (1999).

[11] Chou, J. S., Nguyen, T. K. Forward Forecast of Stock Price Using Sliding-Window Metaheuristic-Optimized Machine-Learning Regression. IEEE Transactions on Industrial Informatics, 14 (7):3132-3142 (2018).

[12] Yang, Q., Wang, C. A study on forecast of global stock indices based on deep LSTM neural networks. Statistical Research, 36 (3):65-78 (2019). https://doi.org/10.19343/j.cnki.11-1302/c.2019.03.006

[13] Xu, Q. H. A Transformer stock price prediction model based on CNN-GRU. Master's thesis, Jiangxi University of Finance and Economics (2025). CNKI.

[14] Fama, E. F. Efficient capital markets: A review of theory and empirical work. Journal of Finance, 25 (2):383–417 (1970).

[15] Kahneman, D., Tversky, A. Prospect theory: An analysis of decision under risk. Econometrica, 47 (2):263–291 (1979).