Mixture density conditional generative adversarial network models (MD-cGAN)
Generative Adversarial Networks (GANs) have gained significant attention in recent years, with impressive applications highlighted in computer vision, in particular. Compared to such examples, however, there have been more limited applications of GANs to time series modeling, including forecasting....
| Main Authors: | , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
MDPI
2021
|
| _version_ | 1811139597873184768 |
|---|---|
| author | Zand, J Roberts, S |
| author_facet | Zand, J Roberts, S |
| author_sort | Zand, J |
| collection | OXFORD |
| description | Generative Adversarial Networks (GANs) have gained significant attention in recent years, with impressive applications highlighted in computer vision, in particular. Compared to such examples, however, there have been more limited applications of GANs to time series modeling, including forecasting. In this work, we present the Mixture Density Conditional Generative Adversarial Model (MD-CGAN), with a focus on time series forecasting. We show that our model is capable of estimating a probabilistic posterior distribution over forecasts and that, in comparison to a set of benchmark methods, the MD-CGAN model performs well, particularly in situations where noise is a significant component of the observed time series. Further, by using a Gaussian mixture model as the output distribution, MD-CGAN offers posterior predictions that are non-Gaussian. |
| first_indexed | 2024-09-25T04:08:37Z |
| format | Journal article |
| id | oxford-uuid:38dbf9f8-ef78-4792-a54c-3668c5724ce6 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-09-25T04:08:37Z |
| publishDate | 2021 |
| publisher | MDPI |
| record_format | dspace |
| spelling | oxford-uuid:38dbf9f8-ef78-4792-a54c-3668c5724ce62024-06-11T14:37:13ZMixture density conditional generative adversarial network models (MD-cGAN)Journal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:38dbf9f8-ef78-4792-a54c-3668c5724ce6EnglishSymplectic ElementsMDPI2021Zand, JRoberts, SGenerative Adversarial Networks (GANs) have gained significant attention in recent years, with impressive applications highlighted in computer vision, in particular. Compared to such examples, however, there have been more limited applications of GANs to time series modeling, including forecasting. In this work, we present the Mixture Density Conditional Generative Adversarial Model (MD-CGAN), with a focus on time series forecasting. We show that our model is capable of estimating a probabilistic posterior distribution over forecasts and that, in comparison to a set of benchmark methods, the MD-CGAN model performs well, particularly in situations where noise is a significant component of the observed time series. Further, by using a Gaussian mixture model as the output distribution, MD-CGAN offers posterior predictions that are non-Gaussian. |
| spellingShingle | Zand, J Roberts, S Mixture density conditional generative adversarial network models (MD-cGAN) |
| title | Mixture density conditional generative adversarial network models (MD-cGAN) |
| title_full | Mixture density conditional generative adversarial network models (MD-cGAN) |
| title_fullStr | Mixture density conditional generative adversarial network models (MD-cGAN) |
| title_full_unstemmed | Mixture density conditional generative adversarial network models (MD-cGAN) |
| title_short | Mixture density conditional generative adversarial network models (MD-cGAN) |
| title_sort | mixture density conditional generative adversarial network models md cgan |
| work_keys_str_mv | AT zandj mixturedensityconditionalgenerativeadversarialnetworkmodelsmdcgan AT robertss mixturedensityconditionalgenerativeadversarialnetworkmodelsmdcgan |