Modeling Spatio-Temporal Extreme Events Using Graphical Models

We propose a novel statistical model to describe spatio-temporal extreme events. The model can be used, for instance, to estimate extreme-value temporal pattern such as seasonality and trend, and further to predict the distribution of extreme events in the future. Such model usually involves thousan...

Full description

Saved in:
Bibliographic Details
Main Authors: Yu, Hang, Dauwels, Justin
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2018
Subjects:
Online Access:https://hdl.handle.net/10356/89365
http://hdl.handle.net/10220/44847
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-89365
record_format dspace
spelling sg-ntu-dr.10356-893652020-03-07T14:02:37Z Modeling Spatio-Temporal Extreme Events Using Graphical Models Yu, Hang Dauwels, Justin School of Electrical and Electronic Engineering Extreme Events Graphical Models We propose a novel statistical model to describe spatio-temporal extreme events. The model can be used, for instance, to estimate extreme-value temporal pattern such as seasonality and trend, and further to predict the distribution of extreme events in the future. Such model usually involves thousands or even millions of variables in the spatio-temporal domain, whereas only one single observation is available for each location and time point. To address this challenge, previous works usually employ learning and inference methods that are computationally burdensome, and therefore are prohibitive for large-scale data. Moreover, they assume that the shape and scale parameters of the extreme-value distributions are constant across the spatio-temporal domain, which is often too restrictive in practice. In this paper, we break through these limitations by exploring graphical models to capture the highly structured dependencies among the parameters of extreme-value distributions. Furthermore, we develop an efficient stochastic variational inference (SVI) algorithm to learn the parameters of the resulting non-Gaussian graphical model. The computational complexity of the SVI algorithm is sublinear in the number of variables, thus enabling the proposed model to tackle large-scale spatio-temporal data in real-life applications. Results of both synthetic and real data demonstrate the effectiveness of the proposed approach. MOE (Min. of Education, S’pore) Accepted version 2018-05-21T08:11:27Z 2019-12-06T17:23:55Z 2018-05-21T08:11:27Z 2019-12-06T17:23:55Z 2016 Journal Article Yu, H., & Dauwels, J. (2016). Modeling Spatio-Temporal Extreme Events Using Graphical Models. IEEE Transactions on Signal Processing, 64(5), 1101-1116. 1053-587X https://hdl.handle.net/10356/89365 http://hdl.handle.net/10220/44847 10.1109/TSP.2015.2491882 en IEEE Transactions on Signal Processing © 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: [http://dx.doi.org/10.1109/TSP.2015.2491882]. 16 p. application/pdf
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic Extreme Events
Graphical Models
spellingShingle Extreme Events
Graphical Models
Yu, Hang
Dauwels, Justin
Modeling Spatio-Temporal Extreme Events Using Graphical Models
description We propose a novel statistical model to describe spatio-temporal extreme events. The model can be used, for instance, to estimate extreme-value temporal pattern such as seasonality and trend, and further to predict the distribution of extreme events in the future. Such model usually involves thousands or even millions of variables in the spatio-temporal domain, whereas only one single observation is available for each location and time point. To address this challenge, previous works usually employ learning and inference methods that are computationally burdensome, and therefore are prohibitive for large-scale data. Moreover, they assume that the shape and scale parameters of the extreme-value distributions are constant across the spatio-temporal domain, which is often too restrictive in practice. In this paper, we break through these limitations by exploring graphical models to capture the highly structured dependencies among the parameters of extreme-value distributions. Furthermore, we develop an efficient stochastic variational inference (SVI) algorithm to learn the parameters of the resulting non-Gaussian graphical model. The computational complexity of the SVI algorithm is sublinear in the number of variables, thus enabling the proposed model to tackle large-scale spatio-temporal data in real-life applications. Results of both synthetic and real data demonstrate the effectiveness of the proposed approach.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Yu, Hang
Dauwels, Justin
format Article
author Yu, Hang
Dauwels, Justin
author_sort Yu, Hang
title Modeling Spatio-Temporal Extreme Events Using Graphical Models
title_short Modeling Spatio-Temporal Extreme Events Using Graphical Models
title_full Modeling Spatio-Temporal Extreme Events Using Graphical Models
title_fullStr Modeling Spatio-Temporal Extreme Events Using Graphical Models
title_full_unstemmed Modeling Spatio-Temporal Extreme Events Using Graphical Models
title_sort modeling spatio-temporal extreme events using graphical models
publishDate 2018
url https://hdl.handle.net/10356/89365
http://hdl.handle.net/10220/44847
_version_ 1681046485275246592