Jeremiah Lowin – Machine Learning in Investing – [Invest Like the Best, EP.105]

In this episode of Invest Like the Best, Patrick O'Shaughnessy interviews Jeremiah Lowin about the deployment of machine learning technologies in investing. They discuss the usefulness of machine learning models, describing them as tools for discovering complex correlations in data, and emphasize the importance of understanding their limitations. The conversation covers various aspects of machine learning, including classification, regression, feature engineering, training, testing, and hyperparameter tuning. They also explore the differences between machine learning and classical academic finance, the role of stationarity, and the significance of label formation. The discussion further delves into the challenges and potential of applying machine learning techniques in finance, highlighting the importance of data quality, feature engineering, and the ability of models to "shrug" when uncertain.