Data-driven Hedging with Generative Models: Findings & Methodology

Created using ChatSlide

This coursework introduces data-driven hedging through Generative Adversarial Networks (GANs), emphasizing data’s role in modern strategies. It covers SPX options data collection, cleaning, and building implied volatility surfaces for robust analysis. Key preprocessing techniques are explored, including feature engineering, temporal sequencing, and data splits. Students will analyze VolGAN architecture, loss optimization, and performance metrics. The final module discusses study insights,...