Approximate inference for spatial functional data on massively parallel processors
Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Standard
Approximate inference for spatial functional data on massively parallel processors. / Raket, Lars Lau; Markussen, Bo.
I: Computational Statistics & Data Analysis, Bind 72, 2014, s. 227-240.Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - Approximate inference for spatial functional data on massively parallel processors
AU - Raket, Lars Lau
AU - Markussen, Bo
PY - 2014
Y1 - 2014
N2 - With continually increasing data sizes, the relevance of the big n problem of classical likelihood approaches is greater than ever. The functional mixed-effects model is a well established class of models for analyzing functional data. Spatial functional data in a mixed-effects setting is considered, and so-called operator approximations for doing inference in the resulting models are presented. These approximations embed observations in function space, transferring likelihood calculations to the functional domain. The resulting approximated problems are naturally parallel and can be solved in linear time. An extremely efficient GPU implementation is presented, and the proposed methods are illustrated by conducting a classical statistical analysis of 2D chromatography data consisting of more than 140 million spatially correlated observation points.
AB - With continually increasing data sizes, the relevance of the big n problem of classical likelihood approaches is greater than ever. The functional mixed-effects model is a well established class of models for analyzing functional data. Spatial functional data in a mixed-effects setting is considered, and so-called operator approximations for doing inference in the resulting models are presented. These approximations embed observations in function space, transferring likelihood calculations to the functional domain. The resulting approximated problems are naturally parallel and can be solved in linear time. An extremely efficient GPU implementation is presented, and the proposed methods are illustrated by conducting a classical statistical analysis of 2D chromatography data consisting of more than 140 million spatially correlated observation points.
U2 - 10.1016/j.csda.2013.10.016
DO - 10.1016/j.csda.2013.10.016
M3 - Journal article
VL - 72
SP - 227
EP - 240
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
SN - 0167-9473
ER -
ID: 74861895