Ensemble Learning Applied to Quant Equity: Gradient Boosting in a Multi-Factor Framework
Big Data and Machine Learning in Quantitative Investment, Wiley finance series. 2018
Posted: 16 Aug 2018
Date Written: February 28, 2018
Abstract
In this chapter, we apply a popular Machine Learning approach (extreme gradient boosted trees) to build enhanced diversified equity portfolios. A simple naïve equally-weighted portfolio of US stocks based on a boosted tree-based signal generates on average an excess return of 3.1% per annum, compared to a simple multifactor portfolio. We demonstrate that using boosted trees on a large number of features give an average error rate of 20% for predicting the 12-month sector neutral outperformance of a stock. In addition, enhancing a simple multi-factor signal with an ML-boosted signal proves to add value on a risk-return basis without altering the factor exposure of the traditional multi-factor portfolio.
Keywords: Machine learning, Factor Investing, Stock selection, Portfolio construction, Quantitative investment
JEL Classification: G11, C61
Suggested Citation: Suggested Citation