Generating financial markets with signatures

Journal article

While most generative models tend to rely on large amounts of training data, here Hans Buehler et al present a generative model that works reliably even in environments where the amount of available training data is small, irregularly paced or oscillatory. They show how a rough paths-based feature map encoded by the signature of the path outperforms returns-based market generation both numerically and from a theoretical point of view. Finally, they propose a suitable performance evaluation metric for financial time series and discuss some connections between their signature-based market generator and deep hedging.

Authors: Buehler, H; Horvath, B; Lyons, T; Perez, I; Wood, B

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Infopro Digital Services Limited
• Journal: Risk.net
• Volume: 2021
• Issue: June
• Article number: 7841726
• Publication date: 2021-06-09
• Acceptance date: 2021-03-11
• ISSN: 1743-9477
• Language: English
• Keywords: FFR; neural networks; signatures; market generator; rough path theory