Some asset managers see machine learning (ML) as a breakthrough for better analysis and prediction. Others argue that these techniques are just specialized tools for quantitative analysts that won’t change basic asset management practices. Machine learning for asset managers, the first in the Cambridge Elements in Quantitative Finance series, is a short book that does not fully answer this big question or serve as a foundational text on the subject. However, it shows how applying the right data analysis techniques can have a significant impact on solving complex asset management problems that cannot be solved by classical statistical analysis.
The traditional approach to the broad subject of machine learning focuses on general prediction techniques and the taxonomy of supervised and unsupervised learning models through the presentation of the differences between machine learning and deep learning, as well as on the major themes of artificial intelligence. (For a traditional general review, cf. Artificial intelligence in asset management by Söhnke M. Bartram, Jürgen Branke, and Mehrshad Motahari.) Marcos M. López de Prado, Chief Investment Officer of True Positive Technologies and Professor of Practice at Cornell University College of Engineering, uses a more modest but compelling approach to present the value of machine learning. This short book will help readers appreciate the potential power of machine learning techniques as it focuses on solutions to tricky asset management problems.
López de Prado’s presentation on problem-solving techniques provides useful insight into machine learning to a wide audience. The primary audience for the book, however, is quantitative analysts who want to learn about new techniques and access Python code that will get them started on their implementation of management solutions. Further analysis can be found in López de Prado’s longer work on the subject, Advances in Financial Machine Learning.
The book’s excellent introduction explains why machine learning techniques will greatly benefit asset managers and why traditional or classic linear techniques have limitations and are often inadequate in asset management. This clearly shows that ML is not a black box but a set of data tools that enhance theory and improve data clarity. López de Prado focuses on seven complex problems or topics where the application of new techniques developed by ML specialists will bring added value.
The first major topic concerns problems with covariance matrices. The noise in the covariance matrix will influence any regression analysis or optimization, so techniques that can better extract signals from noise will improve portfolio management decisions. The second topic in this same general area shows how to “detonate” the covariance matrix by extracting the market component that often floods other valuable information from the covariance matrix. The expansion of techniques for extracting data signals will enable better asset management decisions.
Next, López de Prado discusses how the distance matrix can be an improved method for looking beyond correlation and how the concept of entropy or co-dependency from information theory can be a useful tool. Building blocks, such as distance functions and clustering techniques, can account for nonlinear effects, nonnormality, and outliers that can unduly influence traditional correlation analysis. For example, optimal clusters can be used to group data of similar quality as an unsupervised learning technique that can effectively provide better insight into the relationships between markets than what is found in the traditional correlation matrix. .
For those interested in the central problem of prediction, López de Prado addresses the often overlooked topic of financial labeling – that is, setting forecasting goals as a key learning problem. supervised. Horizon returns are neither the only nor the best method of labeling data for forecasts. For example, most traders are not interested in the difficult problem of predicting a point estimate of where a stock will be in a week or month. They are, however, very interested in a model that accurately predicts the direction of the market. In short, the labels of what is predicted are important.
The book addresses the central problem of p-values and the concept of statistical significance. Attention to this topic has increased within finance due to the “zoo” of statistically significant risk premia that cannot be replicated out of sample. This discussion demonstrates the wide application of ML as a general tool, not only for problem solving, but also for enhancing theory development. ML techniques such as average decreasing impurity, or MDI, and average decreasing precision, or MDA, can serve as effective and more efficient substitutes for p-values.
Since the innovations of Harry Markowitz, portfolio construction has been a constant source of frustration for asset managers. The “Markowitz curse”, which limits the successful use of optimization when it is most needed, can be solved by using ML techniques such as hierarchical clustering and nested cluster optimization to unravel the relationships of data and simplify the optimal portfolio solution.
The final topic is overfitting testing, a key issue for any quantitative asset manager trying to find that perfect model. ML techniques coupled with Monte Carlo simulations, which use the power of fast computing, can be used to provide multiple backtests and suggest a range of possible Sharpe ratios. A model with a high Sharpe ratio may just be a matter of luck – a return path among a wide range. Using ML can better identify false strategies and the likelihood of Type I or Type II statistical errors. Discovering a failure in the lab will save time and money before strategies are put into production.
Machine learning for asset managers uses color for better display of graphics and contains a significant amount of Python code to help readers who wish to implement the techniques presented. Code snippets are helpful for readers who want to use this research, but sometimes the integration of code and text in this book can be confusing. Although the author is adept at explaining complex topics, some steps, transitions, and conclusions are difficult to follow for anyone lacking in-depth quantitative knowledge. This work mixes up some of the author’s hands-on research projects, but this may be a drawback for readers looking for connections between techniques to think about machine learning holistically.
Brevity is the advantage of this work, but a longer book would better support the author’s attempt to demonstrate how machine learning can facilitate the development of new theories and complement classical statistical theories. For example, the book’s introduction provides one of the best motivations for using machine learning in asset management that I’ve read. In a few short pages, it addresses popular misconceptions, answers frequently asked questions, and explains how machine learning can be directly applied to portfolio management. López de Prado has practical knowledge that most technical writers lack. So it would be helpful for readers to take more advantage of his extensive ML knowledge.
In summary, Machine learning for asset managers successfully demonstrates the power of ML techniques in solving difficult asset management problems, but it should not be considered an introduction to the subject for general asset managers. Still, learning how these techniques can solve problems, as an author who has had significant success in asset management explains, is worth the book’s modest price.
If you liked this article, don’t forget to subscribe to the Enterprising investor.
All posts are the opinion of the author. As such, they should not be construed as investment advice, and the opinions expressed do not necessarily reflect the views of the CFA Institute or the author’s employer.
Professional Learning for CFA Institute Members
CFA Institute members are empowered to self-determine and report professional learning (PL) credits earned, including content on Enterprising investor. Members can easily register credits using their HGV tracker online.