Regression and Forecasting of U.S. Stock Returns Based on LSTM

2025-02-03

AI Summary

This study evaluates the explanatory and predictive power of the Fama–French three- and five-factor models and the Carhart four-factor model for stock returns across U.S. manufacturing, high-technology, and other sectors, while enhancing their capacity to model nonlinear dynamics. Method: We innovatively integrate Long Short-Term Memory (LSTM) networks into the conventional multifactor framework to capture sector-specific temporal nonlinearities; we also conduct the first systematic cross-sector comparison and validation of the five-factor model’s incremental explanatory power. Contribution/Results: Empirical results show that the five-factor model consistently outperforms alternatives across all three sectors. Moreover, LSTM augmentation significantly improves return prediction accuracy—particularly in manufacturing and high-technology stocks—reducing average mean absolute error (MAE) by 18.7%, thereby confirming its effectiveness in capturing dynamic heterogeneities unmodeled by traditional linear factor specifications.

Technology Category

Machine Learning: Large Multimodal Models (LMMs)Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP ModelsData Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphsSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMsSearch and Retrieval-Augmented AI: Large language models for search

Abstract

This paper analyses the investment returns of three stock sectors, Manuf, Hitec, and Other, in the U.S. stock market, based on the Fama-French three-factor model, the Carhart four-factor model, and the Fama-French five-factor model, in order to test the validity of the Fama-French three-factor model, the Carhart four-factor model, and the Fama-French five-factor model for the three sectors of the market. French five-factor model for the three sectors of the market. Also, the LSTM model is used to explore the additional factors affecting stock returns. The empirical results show that the Fama-French five-factor model has better validity for the three segments of the market under study, and the LSTM model has the ability to capture the factors affecting the returns of certain industries, and can better regress and predict the stock returns of the relevant industries. Keywords- Fama-French model; Carhart model; Factor model; LSTM model.

Problem

• Evaluates Fama-French and Carhart models' validity
• Explores additional stock return factors using LSTM
• Predicts stock returns in specific U.S. sectors

Innovation

• LSTM model captures stock return factors
• Fama-French five-factor model validity tested
• Carhart model used for stock analysis