A fast spectral quasi-likelihood approach for spatial point processes

C. Deng, R. P. Waagepetersen, M. Wang, Yongtao Guan

Research output: Contribution to journalArticle

Abstract

In applications of spatial point processes, it is often of interest to fit a parametric model for the intensity function. For this purpose Guan et al. (2015) recently introduced a quasi-likelihood type estimating function that is optimal in a certain class of first-order estimating functions. However, depending on the choice of certain tuning parameters, the implementation suggested in Guan et al. (2015) can be very demanding both in terms of computing time and memory requirements. Using a novel spectral representation, we construct in this paper an implementation that is computationally much more efficient than the one proposed in Guan et al. (2015).

Original languageEnglish (US)
Pages (from-to)59-64
Number of pages6
JournalStatistics and Probability Letters
Volume133
DOIs
StatePublished - Feb 1 2018

Fingerprint

Spatial Point Process
Quasi-likelihood
Estimating Function
Intensity Function
Spectral Representation
Parameter Tuning
Parametric Model
First-order
Computing
Requirements
Point process
Class
Parametric model

Keywords

  • Estimating function
  • Quasi-Likelihood
  • Spatial point process
  • Spectral approach

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

A fast spectral quasi-likelihood approach for spatial point processes. / Deng, C.; Waagepetersen, R. P.; Wang, M.; Guan, Yongtao.

In: Statistics and Probability Letters, Vol. 133, 01.02.2018, p. 59-64.

Research output: Contribution to journalArticle

Deng, C. ; Waagepetersen, R. P. ; Wang, M. ; Guan, Yongtao. / A fast spectral quasi-likelihood approach for spatial point processes. In: Statistics and Probability Letters. 2018 ; Vol. 133. pp. 59-64.
@article{0c97d427609f460d9a4de66fdfd74d4b,
title = "A fast spectral quasi-likelihood approach for spatial point processes",
abstract = "In applications of spatial point processes, it is often of interest to fit a parametric model for the intensity function. For this purpose Guan et al. (2015) recently introduced a quasi-likelihood type estimating function that is optimal in a certain class of first-order estimating functions. However, depending on the choice of certain tuning parameters, the implementation suggested in Guan et al. (2015) can be very demanding both in terms of computing time and memory requirements. Using a novel spectral representation, we construct in this paper an implementation that is computationally much more efficient than the one proposed in Guan et al. (2015).",
keywords = "Estimating function, Quasi-Likelihood, Spatial point process, Spectral approach",
author = "C. Deng and Waagepetersen, {R. P.} and M. Wang and Yongtao Guan",
year = "2018",
month = "2",
day = "1",
doi = "10.1016/j.spl.2017.09.016",
language = "English (US)",
volume = "133",
pages = "59--64",
journal = "Statistics and Probability Letters",
issn = "0167-7152",
publisher = "Elsevier",

}

TY - JOUR

T1 - A fast spectral quasi-likelihood approach for spatial point processes

AU - Deng, C.

AU - Waagepetersen, R. P.

AU - Wang, M.

AU - Guan, Yongtao

PY - 2018/2/1

Y1 - 2018/2/1

N2 - In applications of spatial point processes, it is often of interest to fit a parametric model for the intensity function. For this purpose Guan et al. (2015) recently introduced a quasi-likelihood type estimating function that is optimal in a certain class of first-order estimating functions. However, depending on the choice of certain tuning parameters, the implementation suggested in Guan et al. (2015) can be very demanding both in terms of computing time and memory requirements. Using a novel spectral representation, we construct in this paper an implementation that is computationally much more efficient than the one proposed in Guan et al. (2015).

AB - In applications of spatial point processes, it is often of interest to fit a parametric model for the intensity function. For this purpose Guan et al. (2015) recently introduced a quasi-likelihood type estimating function that is optimal in a certain class of first-order estimating functions. However, depending on the choice of certain tuning parameters, the implementation suggested in Guan et al. (2015) can be very demanding both in terms of computing time and memory requirements. Using a novel spectral representation, we construct in this paper an implementation that is computationally much more efficient than the one proposed in Guan et al. (2015).

KW - Estimating function

KW - Quasi-Likelihood

KW - Spatial point process

KW - Spectral approach

UR - http://www.scopus.com/inward/record.url?scp=85033586445&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85033586445&partnerID=8YFLogxK

U2 - 10.1016/j.spl.2017.09.016

DO - 10.1016/j.spl.2017.09.016

M3 - Article

AN - SCOPUS:85033586445

VL - 133

SP - 59

EP - 64

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

SN - 0167-7152

ER -