Machine Learning in Quantitative Investing: Applications and Research Practice
Summary
This report overview describes the longstanding use of machine learning and artificial intelligence in quantitative investing. It notes that applications were already present during an early-1990s wave of interest, and that use continued in areas such as algorithmic trading and commodity trading advisers despite the computational and algorithmic limits of that period.
The report introduces several practical questions for applying machine learning to secondary-market investing, including common strategy-development errors, performance attribution, recognizing strategy decay, building a machine-learning platform, connecting it to trading systems, and organizing a quantitative hedge fund team. It is presented as the first installment in a series, with later reports intended to explore these subjects in more detail. Two strategies are cited as examples: a short-term equity-index strategy and a longer-term commodity strategy, with Sharpe ratios and annualized returns reported in the source summary. The document excerpt does not provide methods, test periods, implementation details, or caveats behind those figures, so they cannot establish general model performance.
Key ideas
- Machine learning has been applied to quantitative investing since at least the early 1990s.
- Algorithmic trading and commodity trading adviser strategies are cited as areas where machine learning has been used.
- The report raises issues in strategy development, attribution, and identifying when a strategy stops working.
- It also considers infrastructure, trading-system integration, and team organization.
- Two example strategies are summarized, but the excerpt lacks the evidence needed to evaluate their reported results.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.