Formulaic Alpha
Formulaic alpha generation focuses on automatically discovering interpretable mathematical formulas that predict asset price movements, a crucial task in quantitative finance. Current research heavily utilizes deep reinforcement learning, often employing algorithms like REINFORCE, to efficiently search the vast space of possible formulas, with a focus on generating diverse and synergistic sets of alphas rather than individual factors. This research aims to improve the performance and robustness of quantitative trading strategies by creating more reliable and understandable predictive models, impacting both academic understanding of financial markets and practical investment strategies.
Papers
November 5, 2024
September 8, 2024
June 26, 2024
June 24, 2024
January 5, 2024