We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: On Computationally-Scalable Spatio-Temporal Regression Clustering of Precipitation Threshold Excesses

Abstract: Focusing on regression based analysis of extremes in a presence of systematically missing covariates, this work presents a data-driven spatio-temporal regression based clustering of threshold excesses. It is shown that in a presence of systematically missing covariates the behavior of threshold excesses becomes nonstationary and nonhomogenous. The presented approach describes this complex behavior by a set of local stationary Generalized Pareto Distribution (GPD) models, where the parameters are expressed as regression models, and a latent spatio-temporal switching process. The spatio-temporal switching process is resolved by the nonparametric Finite Element Methodology for time series analysis with Bounded Variation of the model parameters (FEM-BV). The presented FEM-BV-GPD approach goes beyond strong a priori assumptions made in standard latent class models like Mixture Models and Hidden Markov Models. In addition, it provides a pragmatic description of the underlying dependency structure. The performance of the framework is demonstrated on historical precipitation data for Switzerland and compared with the results obtained by the standard methods on the same data.
Subjects: Methodology (stat.ME)
Cite as: arXiv:2103.16687 [stat.ME]
  (or arXiv:2103.16687v2 [stat.ME] for this version)

Submission history

From: Olga Kaiser [view email]
[v1] Tue, 30 Mar 2021 21:19:08 GMT (2405kb,D)
[v2] Thu, 1 Apr 2021 16:06:14 GMT (2405kb,D)

Link back to: arXiv, form interface, contact.