File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Multivariate time-series modeling with generative neural networks

TitleMultivariate time-series modeling with generative neural networks
Authors
KeywordsARMA–GARCH model
Copulas
Dependence
Exchange rates
Generative moment matching networks
Learning distributions
Probabilistic forecasts
Yield curves
Issue Date2022
Citation
Econometrics and Statistics, 2022, v. 23, p. 147-164 How to Cite?
AbstractGenerative moment matching networks (GMMNs) are introduced as dependence models for the joint innovation distribution of multivariate time series (MTS). Following the popular copula–GARCH approach for modeling dependent MTS data, a framework based on a GMMN–GARCH approach is presented. First, ARMA–GARCH models are utilized to capture the serial dependence within each univariate marginal time series. Second, if the number of marginal time series is large, principal component analysis (PCA) is used as a dimension-reduction step. Last, the remaining cross-sectional dependence is modeled via a GMMN, the main contribution of this work. GMMNs are highly flexible and easy to simulate from, which is a major advantage over the copula–GARCH approach. Applications involving yield curve modeling and the analysis of foreign exchange-rate returns demonstrate the utility of the GMMN–GARCH approach, especially in terms of producing better empirical predictive distributions and making better probabilistic forecasts.
Persistent Identifierhttp://hdl.handle.net/10722/325543
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorHofert, Marius-
dc.contributor.authorPrasad, Avinash-
dc.contributor.authorZhu, Mu-
dc.date.accessioned2023-02-27T07:34:08Z-
dc.date.available2023-02-27T07:34:08Z-
dc.date.issued2022-
dc.identifier.citationEconometrics and Statistics, 2022, v. 23, p. 147-164-
dc.identifier.urihttp://hdl.handle.net/10722/325543-
dc.description.abstractGenerative moment matching networks (GMMNs) are introduced as dependence models for the joint innovation distribution of multivariate time series (MTS). Following the popular copula–GARCH approach for modeling dependent MTS data, a framework based on a GMMN–GARCH approach is presented. First, ARMA–GARCH models are utilized to capture the serial dependence within each univariate marginal time series. Second, if the number of marginal time series is large, principal component analysis (PCA) is used as a dimension-reduction step. Last, the remaining cross-sectional dependence is modeled via a GMMN, the main contribution of this work. GMMNs are highly flexible and easy to simulate from, which is a major advantage over the copula–GARCH approach. Applications involving yield curve modeling and the analysis of foreign exchange-rate returns demonstrate the utility of the GMMN–GARCH approach, especially in terms of producing better empirical predictive distributions and making better probabilistic forecasts.-
dc.languageeng-
dc.relation.ispartofEconometrics and Statistics-
dc.subjectARMA–GARCH model-
dc.subjectCopulas-
dc.subjectDependence-
dc.subjectExchange rates-
dc.subjectGenerative moment matching networks-
dc.subjectLearning distributions-
dc.subjectProbabilistic forecasts-
dc.subjectYield curves-
dc.titleMultivariate time-series modeling with generative neural networks-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1016/j.ecosta.2021.10.011-
dc.identifier.scopuseid_2-s2.0-85119899540-
dc.identifier.volume23-
dc.identifier.spage147-
dc.identifier.epage164-
dc.identifier.eissn2452-3062-
dc.identifier.isiWOS:000798794100008-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats