Machine Learning Classification and Portfolio Construction: Does the Loss Function Matter?

Title:Machine Learning Classification and Portfolio Construction: Does the Loss Function Matter?

Abstract:Classification outperforms regression across matched machine learning models in portfolio construction. A stacking ensemble of gradient boosted tree, random forest, and neural network yields a value-weighted annualized Sharpe ratio of 1.83 for classification and 1.11 for regression. This outperformance persists in multiclass settings, across subsamples, and after transaction costs. Spanning tests show that classification retains economically large alphas after we control for regression, whereas regression alphas shrink substantially once we control for classification. These results indicate that classification extracts more return information than matched regression. Our diagnostics trace classification's advantage to sharper and more precise separation of return deciles.

General Finance (q-fin.GN); Machine Learning (cs.LG); General Economics (econ.GN); Computational Finance (q-fin.CP); Portfolio Management (q-fin.PM)

arXiv:2108.02283 [q-fin.GN]

(or arXiv:2108.02283v4 [q-fin.GN] for this version)

https://doi.org/10.48550/arXiv.2108.02283 Focus to learn more arXiv-issued DOI via DataCite

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.