CorrGAN: Sampling Realistic Financial Correlation Matrices Using Generative Adversarial Networks
This work addresses the need for realistic correlation matrices in finance, with applications in trading strategies, risk management, and empirical finance, representing a novel contribution as the first documented use of GANs for this purpose.
The authors tackled the problem of sampling realistic financial correlation matrices by proposing a generative adversarial network approach, which successfully recovered most known stylized facts about empirical correlation matrices from asset returns.
We propose a novel approach for sampling realistic financial correlation matrices. This approach is based on generative adversarial networks. Experiments demonstrate that generative adversarial networks are able to recover most of the known stylized facts about empirical correlation matrices estimated on asset returns. This is the first time such results are documented in the literature. Practical financial applications range from trading strategies enhancement to risk and portfolio stress testing. Such generative models can also help ground empirical finance deeper into science by allowing for falsifiability of statements and more objective comparison of empirical methods.