Machine Learning for Asset Managers (Chapter 1)

Abstract

Successful investment strategies are specific implementations of general theories. An investment strategy that lacks a theoretical justification is likely to be false. Hence, an asset manager should concentrate her efforts on developing a theory, rather than on back-testing potential trading rules. The purpose of this monograph is to introduce Machine Learning (ML) tools that can help asset managers discover economic and financial theories.

ML is not a black-box, and it does not necessarily over-fit. ML tools complement rather than replace the classical statistical methods. Some of ML’s strengths include:

(i) Focus on out-of-sample predictability over variance adjudication;

(ii) usage of computational methods to avoid relying on (potentially unrealistic) assumptions;

(iii) ability to “learn” complex specifications, including non-linear, hierarchical and non-continuous interaction effects in a high-dimensional space; and

(iv) ability to disentangle the variable search from the specification search, robust to multicollinearity and other substitution effects.

Keywords: Machine Learning, Unsupervised Learning, Supervised Learning, Clustering, Classification, Labeling, Portfolio Construction

JEL Classification: G0, G1, G2, G15, G24, E44