A Brief Introduction to Applying Machine Learning to Investing - Maximilian Stroh -Dr. Philip Messow

A Brief Introduction to Applying Machine Learning to Investing - Maximilian Stroh -Dr. Philip Messow Applying machine learning to longer-term investing comes with specific difficulties. In this talk, we discuss a number of these issues to be aware of, such as: • Signal-to-noise ratios are low • Markets are continuously evolving • The number of independent observations is lower than you might think • Predictability does not guarantee profitability in real-world portfolios While these problems cannot be circumvented altogether, we review ways to address them at least partially. Furthermore, we present examples where machine learning nevertheless can add value. Examples are from the following two papers: Leung, Lohre, Mischlich, Shae, Stroh: The Promises and Pitfalls of Machine Learning for Predicting Stock Returns, Journal of Financial Data Science, =3546725 Kaufmann, Messow, Vogt: Boosting Momentum, Working Paper, =3668928 SUBSCRIBE: