Machine learning for resource managers. 2020. Marcos M. Lopez de Prado Cambridge University Press (Cambridge Elements in the Quantitative Finance Series).
Some resource managers see machine learning (ML) as a breakthrough for better analysis and prediction. Others argue that these strategies are special tools for quantum analysts that will not change core asset management practices. Machine learning for resource managers, The first in the Cambridge Elements in Quantitative Finance series, a small book that does not fully answer this big question or serve as the basic text of the subject. However, it shows how applying the right data analysis strategy can have a significant impact on solving challenging resource management problems that are not solvable through classical statistical analysis.
The traditional thematic approach to the broader subject matter of machine learning focuses on common prediction techniques and the classification of supervised and unsupervised learning models, as well as the broader theme of artificial intelligence, presenting the differences between machine learning and deep learning. (For a traditional theoretical general review, see Artificial intelligence in asset management Sohke M. Bartram, Jurgen Branke and Mehershad Motahari.) Marcos M. Lopez de Prado, chief investment officer at True Positive Technologies and professor of practice at Cornell University College of Engineering, uses a more modest but compelling approach to presentation. The value of machine learning This brief work will help readers appreciate the potential strengths of machine learning strategies as it focuses on disturbing solutions to resource management problems.
The presentation of Lopez de Prado’s problem-solving strategies provides a useful taste of machine learning for a wide audience. The book’s initial audience, however, consists of quantitative analysts who want to read about new strategies and access Python code that will speed up the implementation of their management solutions. A more in-depth analysis can be found of Lopez de Prado’s long work on the subject, Progress in Financial Machine Learning.
The excellent introduction to the book explains why machine learning techniques will benefit resource managers considerably and why traditional theoretical or classical linear techniques have limitations and are often inadequate in resource management. This makes a strong case that ML is not a black box but a set of data tools that improve theory and improve data clarity. López de Prado focuses on seven complex problems or issues where applying new techniques invented by ML experts will increase value.
The first major issue involves problems with the intercourse matrix. Noise in the covariance matrix will affect any regression analysis or optimization, so strategies that can better extract signals from noise will improve portfolio management decisions. This second issue in the same general area shows how to “detonate” the covariance matrix by extracting the market component that is often embedded in other valuable covariance matrix information. Expanding the techniques of data signal extraction will support better decision making of asset management.
Later, Lopez de Prado explained how the matrix of distances could be an advanced method for looking outside the interrelationships and how the concept of entropy or code dependence from information theory could be a useful tool. Building blocks, such as distance functions and clustering techniques, may be responsible for nonlinear effects, abnormalities, and outsiders that may unreasonably influence traditional thematic correlation analysis. For example, optimal clusters can be used to group similar quality data as missing learning strategies that can provide more insights into relationships across the market than conventional correlation matrix.
For those interested in the key issue of prediction, Lopez de Prado discusses the often overlooked issue of financial labeling – that is, the purpose of forecasting as an important topic in supervised learning. Horizon returns are not the only or best way to label data for prediction. For example, most traders are not interested in the difficult problem of predicting a point where a stock will be in a week or a month. They are very interested, however, in a model that accurately directs the market. In short, labels are important for what is being predicted.
The book solves the main problem P-The concept of value and statistical significance. Statistically significant risks are increasing due to premium “zoo” attention to this issue that cannot be replicated outside the sample. This discussion shows the widespread application of ML as a general tool, not only for problem solving, but also for the improvement of theory. This type of ML technique can serve as an effective and more efficient alternative to impurity, or MDI, and reduction accuracy, or MDA. P-Price.
Since the invention of Harry Markovitz, portfolio construction has been a source of ongoing frustration for asset managers. The “Markovits Curse”, which limits the successful use of optimization when needed, can be solved using ML techniques such as sequential clustering and nested cluster optimization to stimulate data relationships and simplify optimal portfolio solutions.
The final issue is testing for overfitting, a key issue for any quantitative asset manager trying to find the perfect model. ML techniques combined with Monte Carlo simulations, which use the power of fast computing, can be used to provide multiple backtests and offer a range of possible sharp ratios. A model with a high sharpness ratio can be just a matter of luck – a return path from a wide range. False techniques using ML and the possibility of type I or type II statistical errors can be better identified. Discovering failure in the laboratory will save time and money before producing the technique.
Machine learning for resource managers There is a significant amount of Python code to use color for better display graphics and to help readers who want to implement the presented strategy. Code snippets are useful 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 issues, it is difficult for anyone who lacks some knowledge to follow certain steps, transformations, and conclusions. This work mixes up some of the author’s practical research projects, but it can be a hassle for readers to find connections between strategies for thinking about machine learning.
Conciseness is the advantage of this work, but a long book will better support the author’s efforts on how machine learning can facilitate the development of new theories and complement the classical statistical theory. For example, the introduction to the book I read provides one of the best motivations to use machine learning in asset management. In just a few short pages, it solves popular misconceptions, answers frequently asked questions and explains how machine learning can be applied directly to portfolio management. Lopez de Prado has practical insights that most technical writers lack, so drawing more extensively on his deep ML knowledge will be helpful for readers.
In short, Machine learning for resource managers The success of ML strategies in solving difficult asset management problems shows success, but for general asset managers this should not be seen as an introduction to the issue. Nevertheless, learning how these techniques can solve problems, such as an author who has enjoyed significant success in asset management, explains the general value of the book.
If you liked this post, be sure to subscribe Entrepreneurial investors.
All posts are the author’s opinion. As such, they should not be construed as investment advice, or the opinions expressed must not reflect the views of the CFA Institute or the author’s employer.
Professional education for CFA Institute members
Members of the CFA Institute have the ability to earn self-assessment and self-reporting professional education (PL) credit, including content Entrepreneurial investors. Members can easily record credit using their online PL tracker.