Machine Learning, News Analytics, and Stock Selection

Yin Luo, Managing Director, Global Head of Quantitative Strategy, Deutsche Bank.
Big data and machine learning have generated tremendous interest in empirical finance research. In this paper, we study a unique news analytics database provided by Ravenpack. We apply a suite of innovative machine learning algorithms, including adaBoost, spline regression, and other boosting/bagging techniques on both traditional and unstructured news data in predicting stock returns. We find news sentiment data adds significant incremental predictive power to our machine learning based global stock selection models.
Session recorded June 16 2016 at the RavenPack 4th Annual Research Conference, titled “Reshaping Finance with Alternative Data”.