EcoSta 2022: Start Registration
View Submission - EcoSta2022
A0284
Title: Cross-sectional analysis of conditional stock returns: quantile regression with machine learning Authors:  Haitao Li - Cheung Kong Graduate School of Business (China)
Guoliang Ma - Iowa State University (United States)
Cindy Yu - Iowa State University (United States) [presenting]
Abstract: Machine learning methods are developed to forecast conditional quantiles of stock returns in the cross section through quantile regression. Machine learning makes it possible to capture highly nonlinear relations between conditional quantiles and a large number of return predictors. We adopt Bayesian optimization with Gaussian process that significantly improves the efficiency of hyperparameter tuning in machine learning. Simulation studies show that our methods accurately predict the conditional quantiles and consequently the whole conditional distributions of complicated data-generating processes. Empirical results show that our methods can identify stocks with extreme positive or negative returns and achieve superior performance in long-short investing.