Application of AdaBoost in Predicting Individual Stock Returns

DOI

10.2991/978-94-6239-801-6_46How to use a DOI?

Keywords

AdaBoost; Individual Stock Returns; Integrated Learning

Abstract

Since the stock market is affected by multiple factors, stock return series exhibit significant nonlinear and nonstationary characteristics, and traditional linear models struggle to effectively fit their volatility patterns. This paper takes the daily return of a share as the research object, constructs a simplified factor system covering volume and price, technology, momentum, and nonlinear interaction factors, and constructs three models of linear regression, random forest, and AdaBoost for comparative analysis. Focusing on the problem that the AdaBoost model is easily disturbed by redundant factors and has weak generalization ability in regression tasks, special optimization is carried out from three aspects: factor screening, parameter adjustment, and outlier processing. The empirical results show that the optimized AdaBoost model is significantly better than linear regression in the test set, and is close to the performance level of a random forest. Factor importance analysis shows that the momentum factor and the volatility factor contribute the most to yield forecasting, and the nonlinear interaction factor also plays a positive role. The research shows that through extreme factor reduction and parameter optimization, the AdaBoost model can effectively improve the prediction accuracy of individual stock returns, which can provide a reference for quantitative investment decision-making.

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  - Shuotong Xing
PY  - 2026
DA  - 2026/10/23
TI  - Application of AdaBoost in Predicting Individual Stock Returns
BT  - Proceedings of the 2026 2nd International Conference on Financial Innovation and Marketing Management (FIMM 2026)
PB  - Atlantis Press
SP  - 425
EP  - 432
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-801-6_46
DO  - 10.2991/978-94-6239-801-6_46
ID  - Xing2026
ER  -