2019
DOI: 10.3390/e21020184
|View full text |Cite
|
Sign up to set email alerts
|

Bayesian Recurrent Neural Network Models for Forecasting and Quantifying Uncertainty in Spatial-Temporal Data

Abstract: Recurrent neural networks (RNNs) are nonlinear dynamical models commonly used in the machine learning and dynamical systems literature to represent complex dynamical or sequential relationships between variables. More recently, as deep learning models have become more common, RNNs have been used to forecast increasingly complicated systems. Dynamical spatio-temporal processes represent a class of complex systems that can potentially benefit from these types of models. Although the RNN literature is expansive a… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
43
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
7

Relationship

0
7

Authors

Journals

citations
Cited by 87 publications
(43 citation statements)
references
References 67 publications
0
43
0
Order By: Relevance
“…Such models can be implemented in some contexts (e.g. Chatzis, 2015;Chien and Ku, 2016;Gan et al, 2016;McDermott and Wikle, 2017a) but are quite sensitive to particular data sets and are typically computationally prohibitive. More recently, approximate Bayesian methods such as variational Bayes (Tran et al, 2018), and scalable Bayesian methods (Snoek et al, 2015) have been used successfully in deep models.…”
Section: Combining the Dh-dstm And Dn-dstm Frameworkmentioning
confidence: 99%
See 1 more Smart Citation
“…Such models can be implemented in some contexts (e.g. Chatzis, 2015;Chien and Ku, 2016;Gan et al, 2016;McDermott and Wikle, 2017a) but are quite sensitive to particular data sets and are typically computationally prohibitive. More recently, approximate Bayesian methods such as variational Bayes (Tran et al, 2018), and scalable Bayesian methods (Snoek et al, 2015) have been used successfully in deep models.…”
Section: Combining the Dh-dstm And Dn-dstm Frameworkmentioning
confidence: 99%
“…McDermott and Wikle (2017b) made several modifications to the standard ESN model to account for a simple approach to uncertainty quantification in a spatio-temporal nonlinear forecasting setting. They considered a quadratic ESN model.…”
Section: An Ensemble Approachmentioning
confidence: 99%
“…Zhang et al [14] adopted LSTM to predict Sea Surface Temperature (SST). Mcdermott et al [15,16] made simple modifications to the basic RNN and applied it to a Lorenz simulation and two real-world nonlinear spatio-temporal forecasting applications. Moreover, they built a quadratic echo state network to do ensemble forecasting.…”
Section: Introductionmentioning
confidence: 99%
“…There have also been Bayesian implementations of the ESN model (e.g., Chatzis, ; Li, Han, & Wang, ), but none of these has been implemented within a Markov chain Monte Carlo (MCMC) framework, as is the case here, where multiple levels of uncertainties can be accounted for. In the context of spatio‐temporal forecasting, McDermott and Wikle (, ) have shown that augmenting the traditional ESN with quadratic output terms (analogous to the quadratic nonlinear component in statistical DSTMs) and input embeddings (e.g., including lags of the input variables as motivated by Takens's () representation in dynamical systems) can improve forecast accuracy compared to traditional ESN models.…”
Section: Introductionmentioning
confidence: 99%