|Publication type:||Article in scientific journal|
|Type of review:||Peer review (publication)|
|Title:||Tackling the exponential scaling of signature-based generative adversarial networks for high-dimensional financial time-series generation|
|Authors:||De Meer Pardo, Fernando|
|Published in:||The Journal of Financial Data Science|
|Publisher / Ed. Institution:||Portfolio Management Research|
|Subjects:||GAN; Hierarchial clustering; Overfitting; Portfolio construction|
|Subject (DDC):||332: Financial economics|
|Abstract:||Generative adversarial networks (GANs) have been shown to be able to generate samples of complex financial time series, particularly by employing the concept of path signatures, a universal description of the geometric properties of a data stream whose expected value uniquely characterizes the time series. Specifically, the SigCWGAN model (Ni et al. 2020) can generate time series of arbitrary length; however, the parameters of the neural network employed grow exponentially with the dimension of the underlying time series, which makes the model intractable when seeking to generate large financial market scenarios. To overcome this problem of dimensionality, the authors propose an iterative generation procedure relying on the concept of hierarchies in financial markets. The authors construct an ensemble of GANs that they call the Hierarchical-SigCWGAN, which is based on hierarchical clustering that approximates signatures in the spirit of the original model. The Hierarchical-SigCWGAN can scale to higher dimensions and generate large-dimensional scenarios in which the joint behavior of all the assets in the market is replicated. The model is validated by comparing its performance on a series of similarity metrics with respect to the original SigCWGAN on a dataset in which it is still tractable and by showing its scalability on a larger dataset.|
|License (according to publishing contract):||Licence according to publishing contract|
|Departement:||School of Engineering |
School of Management and Law
|Organisational Unit:||Institute of Data Analysis and Process Design (IDP) |
Institute of Wealth & Asset Management (IWA)
|Appears in collections:||Publikationen School of Engineering|
Publikationen School of Management and Law
Files in This Item:
There are no files associated with this item.
Show full item record
De Meer Pardo, F., Schwendner, P., & Wunsch, M. (2022). Tackling the exponential scaling of signature-based generative adversarial networks for high-dimensional financial time-series generation. The Journal of Financial Data Science, 4(4), 110–132. https://doi.org/10.3905/jfds.2022.1.109
De Meer Pardo, F., Schwendner, P. and Wunsch, M. (2022) ‘Tackling the exponential scaling of signature-based generative adversarial networks for high-dimensional financial time-series generation’, The Journal of Financial Data Science, 4(4), pp. 110–132. Available at: https://doi.org/10.3905/jfds.2022.1.109.
F. De Meer Pardo, P. Schwendner, and M. Wunsch, “Tackling the exponential scaling of signature-based generative adversarial networks for high-dimensional financial time-series generation,” The Journal of Financial Data Science, vol. 4, no. 4, pp. 110–132, 2022, doi: 10.3905/jfds.2022.1.109.
De Meer Pardo, Fernando, et al. “Tackling the Exponential Scaling of Signature-Based Generative Adversarial Networks for High-Dimensional Financial Time-Series Generation.” The Journal of Financial Data Science, vol. 4, no. 4, 2022, pp. 110–32, https://doi.org/10.3905/jfds.2022.1.109.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.