Statistical analysis of Wasserstein GANs with applications to time series forecasting
arXiv:2011.03074
Abstract
We provide statistical theory for conditional and unconditional Wasserstein generative adversarial networks (WGANs) in the framework of dependent observations. We prove upper bounds for the excess Bayes risk of the WGAN estimators with respect to a modified Wasserstein-type distance. Furthermore, we formalize and derive statements on the weak convergence of the estimators and use them to develop confidence intervals for new observations. The theory is applied to the special case of high-dimensional time series forecasting. We analyze the behavior of the estimators in simulations based on synthetic data and investigate a real data example with temperature data. The dependency of the data is quantified with absolutely regular beta-mixing coefficients.
arXiv comment47 pages, 4 figures
References in corpus (12)
- Conditional Generative Adversarial Nets
- Improved Training of Wasserstein GANs
- Basic Properties of Strong Mixing Conditions. A Survey and Some Open Questions
- Concentration around the mean for maxima of empirical processes
- T-CGAN: Conditional Generative Adversarial Network for Data Augmentation in Noisy Time Series with Irregular Sampling
- Sorting out Lipschitz function approximation
- Conditional GAN for timeseries generation
- Some Theoretical Insights into Wasserstein GANs
- Some Theoretical Properties of GANs
- On the Minimax Optimality of Estimating the Wasserstein Metric
- Forecasting time series with encoder-decoder neural networks
- Conditional WGANs with Adaptive Gradient Balancing for Sparse MRI Reconstruction