Investor Sentiment and Stock Return Prediction: Evidence from Machine Learning and VAR Models

DOI

10.2991/978-94-6239-801-6_47How to use a DOI?

Keywords

Investor Sentiment; Machine Learning; Stock Return Prediction; Vector Autoregression; Sentiment Dynamics Analysis

Abstract

In the context of rapid information technology development and improved data acquisition capabilities, traditional linear-assumption-based asset pricing models gradually face challenges in short-term return forecasting. As an important behavioral factor affecting stock prices, investor sentiment is of increasing concern about its role in irrational volatility. Taking China’s A-share market as the research object, this paper constructs a multi-dimensional factor system covering fundamentals, trading characteristics, and investor sentiment within a unified data framework, and applies a multifactor linear model, a random forest model, and an XGBoost model to predict stock returns separately. At the same time, the vector autoregressive model is introduced to characterize the dynamic relationship between investor sentiment and stock returns. Empirical results show that machine learning models outperform traditional linear models in out-of-sample prediction and portfolio performance, with the XGBoost model performing the best. After the inclusion of sentiment factors, the predictive power and risk-adjusted returns of the models are significantly improved. Further analysis shows that investor sentiment has a significant but short-term impact on stock returns. The results of this paper show that investor sentiment has a certain explanatory power in short-term earnings forecasting and improves the performance of the model to a certain extent.

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  - Zixuan Yu
PY  - 2026
DA  - 2026/10/23
TI  - Investor Sentiment and Stock Return Prediction: Evidence from Machine Learning and VAR Models
BT  - Proceedings of the 2026 2nd International Conference on Financial Innovation and Marketing Management (FIMM 2026)
PB  - Atlantis Press
SP  - 433
EP  - 441
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-801-6_47
DO  - 10.2991/978-94-6239-801-6_47
ID  - Yu2026
ER  -