期刊
JOURNAL OF EMPIRICAL FINANCE
卷 67, 期 -, 页码 288-317出版社
ELSEVIER
DOI: 10.1016/j.jempfin.2022.04.001
关键词
Stock return; Out-of-sample performance; Combination forecast; Machine learning; Stacking
资金
- National Natural Science Foundation of China [72173068, 71803091]
- MOE (Ministry of Education in China) Project of Humanities and Social Sciences [18YJC790015]
- Social Science Development Fund of Nankai University
- Fundamental Research Funds for the Central Universities
In this study, we utilize the ensemble learning approach of stacking to refine and combine various linear and nonlinear models for stock return prediction. We find that stacking with a simple structure outperforms traditional benchmarks and other models, both in terms of in- and out-of-sample performance measures, particularly during extreme market downturns.
We employ an ensemble learning approach, stacking, to refine and combine a variety of linear and nonlinear individual stock return prediction models. In an application of forecasting U.S. market excess return, stacking with a simple structure can outperform the traditional historical mean benchmark, Mallows model averaging, simple combination forecast, complete subset regression, combination elastic net forecast, and several other models in terms of both in- and out-of-sample performance measures on a consistent basis. More importantly, we find that the out-of-sample gains of stacking are especially evident during extreme downside market movements. Overall, stacking can generate substantive improvements in market excess return predictability.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据