In what areas does machine learning have advantages over conventional financial/investment analysis? In his June 2018 presentation entitled “Nine Financial Applications of Machine Learning”, Marcos Lopez de Prado summarizes investing-related areas in which well-supervised machine learning outperforms conventional methods. Based on relevant research and his experience, he asserts that:
- Machine learning beats linear factor models for price prediction because it focuses on price rather than factor selection and factor coefficients.
- Similarly, machine learning is effective in controlling for a potentially confounding variable directly rather than via an assumption of linear effect and coefficient.
- Portfolios constructed via machine learning beat equal weight and mean-variance optimization (which does not beat equal weight).
- Machine learning is particularly effective at identifying regime changes via cluster analysis.
- Machine learning is effective at position sizing via two-step analysis that first determines direction of trade and then sets the size of the trade.
- Machine learning directly relates a large number of data features to outcomes, thereby exposing which features are most important. Essentially, machine learning sacrifices closed-form specifications (theoretical formulas) for breadth of variables considered.
- Machine learning algorithms can replicate, and may be able to predict, a large percentage of analyst and credit rating agency recommendations.
- Machine learning is well-suited to extraction of sentiment via natural language processing.
In summary, well-supervised machine learning is often more effective for investment analysis than conventional modeling because it accesses a large number of data features without restrictive assumptions.
Whether they have access to machine learning or not, investors are increasingly in competition with it.
Cautions regarding assertions include:
- Application of machine learning requires compute power, sophisticated software and skilled supervision. These capabilities are costly and beyond the reach of most investors, who would bear fees for delegating them to a fund manager.
- As noted in the presentation:
- Overlapping data confounds determination of the connection between outcome and data features.
- Financial series may have properties inconsistent with machine learning assumptions, promoting discovery of patterns when there are none.
Frequently Asked Questions
What is machine learning in investing and how is it different from conventional financial analysis?
Machine learning in investing is the use of algorithms that directly relate a large number of data features to outcomes, such as prices, regimes, or sentiment, without restrictive assumptions about functional form. Marcos Lopez de Prado argues that, when well supervised, it sacrifices closed-form theoretical specifications for the breadth of variables it can consider. Conventional methods, by contrast, typically rely on assumptions such as linear factor effects and fixed coefficients.
Does machine learning outperform traditional methods for portfolio construction and price prediction?
According to Marcos Lopez de Prado's June 2018 presentation, well-supervised machine learning often outperforms conventional methods in investment analysis. He asserts that machine-learning portfolios beat equal weight and mean-variance optimization (which does not beat equal weight), and that machine learning beats linear factor models for price prediction. These are conceptual assertions based on research and his experience, not a new empirical test.
How does machine learning identify market regime changes and set position sizes?
Machine learning identifies regime changes in markets primarily through cluster analysis, which Lopez de Prado describes as particularly effective for this task. For position sizing, it uses a two-step analysis that first determines the direction of a trade and then sets the size of the trade. It can also control for a confounding variable directly rather than assuming a linear effect and coefficient.
Is machine learning better than mean-variance optimization and linear factor models?
Per Lopez de Prado, well-supervised machine learning is often more effective than mean-variance optimization and linear factor models because it accesses many data features without restrictive assumptions. He states that machine-learning portfolios beat equal weight and mean-variance optimization, and that machine learning focuses on price itself rather than factor selection and coefficients. The comparison rests on his assertions and experience, and applies when the machine learning is well supervised.
What are the risks and limitations of using machine learning for investing?
Machine learning for investing requires compute power, sophisticated software and skilled supervision, which are costly and beyond the reach of most investors, who would bear fees for delegating them to a fund manager. Overlapping data confounds determination of the connection between outcomes and data features. Financial series may also have properties inconsistent with machine learning assumptions, promoting discovery of patterns when there are none.