Application of LSTM Models in Stock Return Forecasting

DOI

10.2991/978-94-6239-801-6_38How to use a DOI?

Keywords

Stock Return; Earnings Forecasts; LSTM Model

Abstract

With the accelerated volatility of the stock market and significant differentiation of individual stock returns in recent years, accurate prediction of individual stock returns has become increasingly crucial for financial investment. However, practical issues remain regarding the stability of existing prediction models, necessitating further exploration of reliable return prediction tools to meet market needs. This paper focuses on the application of LSTM models in individual stock return prediction, aiming to capture nonlinear characteristics for more accurate forecasts. It first defines concepts of LSTM, ARIMA, and other models, and reviews their existing applications as a theoretical basis. Subsequently, the LSTM model is constructed, explanatory variables are selected, and test results are visually presented. Finally, robustness evaluation indicators such as RMSE and MAE are adopted to analyze the model’s fitting effect. The study finds that LSTM can fully account for nonlinear time-series dependencies in individual stock returns and achieve certain results in capturing complex market patterns, but it also has shortcomings, including low direction prediction accuracy and training efficiency. Thus, increasing model capacity and selecting more characteristic data is recommended to improve prediction accuracy.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yanning Li
PY  - 2026
DA  - 2026/10/23
TI  - Application of LSTM Models in Stock Return Forecasting
BT  - Proceedings of the 2026 2nd International Conference on Financial Innovation and Marketing Management (FIMM 2026)
PB  - Atlantis Press
SP  - 340
EP  - 347
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-801-6_38
DO  - 10.2991/978-94-6239-801-6_38
ID  - Li2026
ER  -