Ensemble learning and stock return predictability

中央财经大学金融学院副教授

Many, even sophisticated, models cannot beat a simple mean combination of univariate stock market return forecasts. We introduce an ensemble learning method, which averages forecasts from sophisticated models (like BMA, WALS and Lasso) based on random subsamples and adaptively changes sampling distributions. Empirically, our novel method seems to improve the simple mean forecast with statistica.