No. 259 of 500 No. 21 in AI by Industry

Advances in Financial Machine Learning

By Marcos López de Prado

Print length
400 pages
Publisher
Wiley
Reading level
Advanced
ISBN-13
978-1119482086
Published
February 21, 2018
Written for
Quantitative researchers, data scientists and portfolio managers building systematic trading or investment strategies.

Synopsis

Marcos López de Prado wrote this book for quantitative researchers and portfolio managers who want to apply machine learning to investing without fooling themselves. His starting point is that financial data violate the assumptions behind most textbook methods: observations overlap in time, signals are weak, series are nonstationary, and researchers test so many ideas that false discoveries are almost guaranteed. The book's five parts work through these problems in order, covering how to sample market data into information-driven bars, label outcomes with the triple-barrier method and meta-labeling, weight overlapping samples, and use fractional differentiation to make series stationary while preserving memory.

Later chapters adapt ensemble methods, cross-validation and feature importance to finance, introducing purged and embargoed cross-validation to stop information leaking from test sets into training. A long section on backtesting explains why so many backtested strategies fail out of sample and offers tools such as combinatorial purged cross-validation and deflated performance statistics, followed by chapters on structural breaks, entropy and market microstructure features, and high-performance computing. Python snippets accompany the ideas, but this is a research methodology book rather than a coding tutorial. Readers gain a vocabulary and set of safeguards that have become common reference points in quantitative finance.

Scorecard

Editorial score

How it scored

Editorial score: 76.5 / 1005 weighted criteria

  • Influence30%

    24.0/30
  • Authority25%

    22.5/25
  • Practical value20%

    15.0/20
  • Currency15%

    10.5/15
  • Readability10%

    4.5/10
  • Editorial score

    76.5/100

How the rank was computed

Rank No. 259 of 500 comes from a composite of 54.5: observable signals 20.1 of 55 points plus editorial review 34.4 of 45 points.

All 11 inputs
InputPoints
Observable signals20.1/55
Reach (Amazon ratings volume)10.4/16.5
Consensus across best-of lists2.8/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted5.5/8.3
Amazon sales rank1.4/5.5
Recency0.0/2.8
Editorial review34.4/45
Author authority10.1/11.3
Influence10.8/13.5
Practical value6.8/9.0
Readability2.0/4.5
Currency4.7/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A demanding, influential manual that explains why standard machine learning fails on market data and how to fix it.

Strengths

  • Addresses finance-specific pitfalls generic ML texts ignore
  • Rigorous treatment of backtest overfitting
  • Original methods from a leading practitioner

Keep in mind

Dense mathematics, terse code snippets and an assumed background in quantitative finance make it hard going for newcomers and casual investors.

Key takeaways

  1. Time bars sample markets poorly; volume, dollar and information-driven bars produce series with better statistical properties for machine learning.
  2. The triple-barrier method labels trades by whichever is hit first among a profit target, a stop loss and a time limit.
  3. Standard k-fold cross-validation leaks information when financial labels overlap, so purging and embargoing training samples yields more reliable performance estimates.
  4. Every backtest variation is a hypothesis test, so researchers must record all trials and deflate reported Sharpe ratios for multiple testing.

Notable facts

  • The Journal of Portfolio Management named López de Prado its 2019 Quant of the Year, and the announcement highlights this book's proposal to use machine learning to help discover economic theories rather than make black-box predictions. Source (opens in a new tab)

About the author

Marcos López de Prado

Marcos López de Prado is Global Head of Quantitative R&D at the Abu Dhabi Investment Authority, a professor of practice at Cornell's College of Engineering and a research fellow at Lawrence Berkeley National Laboratory. When the book appeared he led machine learning at AQR Capital Management. The Journal of Portfolio Management named him its 2019 Quant of the Year. He holds two doctorates from Universidad Complutense de Madrid.

Questions

What is Advances in Financial Machine Learning about?
It explains how to apply machine learning to investment research while avoiding the errors that cause many strategies to fail. Marcos López de Prado covers structuring market data, labeling outcomes, weighting overlapping samples, finance-specific cross-validation, feature importance, backtest overfitting and high-performance computing, pairing each problem with a mathematically grounded fix and Python code.
Is Advances in Financial Machine Learning good for beginners?
No. The book assumes comfort with statistics, Python and the mechanics of financial markets, and it moves quickly through advanced material. Beginners should first study general machine learning and basic quantitative finance, then return to this book once they are ready to build and test systematic strategies.
Is Advances in Financial Machine Learning still relevant in 2026?
Yes. Its central warnings about nonstationary data, leakage in cross-validation and backtest overfitting apply to any model used for trading, including deep learning and language model signals. Some code and computing chapters show their 2018 vintage, and the author has since extended these ideas in later books, but the methodology remains a standard reference among quantitative researchers.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.