A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data

Michael Renfrew; Bruce J Worton

doi:10.1007/978-3-031-65723-8_20

A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data

Michael Renfrew, Bruce J Worton

School of Mathematics

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

Abstract

The paper compares two advanced boosting techniques, Extreme Gradient Boosting (XGBoost) and Gaussian Process Boosting (GPBoost), for spatial logistic regression models using satellite data. It highlights the limitations of XGBoost in handling autocorrelated data typical in spatial statistics and demonstrates how GPBoost can incorporate a Gaussian process within a mixed-effects model to significantly reduce generalization error.

Original language	English
Title of host publication	Developments in Statistical Modelling
Subtitle of host publication	IWSM 2024
Editors	J Einbeck, H Maeng, E Ogundimu, K Perrakis
Publisher	Springer
Pages	128-133
ISBN (Electronic)	978-3-031-65723-8
ISBN (Print)	978-3-031-65722-1
DOIs	https://doi.org/10.1007/978-3-031-65723-8_20
Publication status	Published - 12 Jul 2024

Publication series

Name	Developments in Statistical Modelling

Access to Document

10.1007/978-3-031-65723-8_20

Renfrew_and_Worton_Accepted_Manuscript
This version of the contribution has been accepted for publication, after peer review (when applicable), but it is not the Version of Record and does not reflect post-acceptance improvements or corrections. The Version of Record is available online at: https://doi.org/10.1007/978-3-031-65723-8_20. Use of this Accepted Version is subject to the publisher’s Accepted Manuscript terms of use, which can be found here: https://www.springernature.com/gp/open-research/policies/accepted-manuscript-terms.
Accepted author manuscript, 305 KBLicence: All Rights Reserved

Cite this

Renfrew, M., & Worton, B. J. (2024). A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data. In J. Einbeck, H. Maeng, E. Ogundimu, & K. Perrakis (Eds.), Developments in Statistical Modelling: IWSM 2024 (pp. 128-133). (Developments in Statistical Modelling). Springer. https://doi.org/10.1007/978-3-031-65723-8_20

@inbook{c01da23824d0430c8cfa9a9f0e3b6883,

title = "A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data",

abstract = "The paper compares two advanced boosting techniques, Extreme Gradient Boosting (XGBoost) and Gaussian Process Boosting (GPBoost), for spatial logistic regression models using satellite data. It highlights the limitations of XGBoost in handling autocorrelated data typical in spatial statistics and demonstrates how GPBoost can incorporate a Gaussian process within a mixed-effects model to significantly reduce generalization error.",

author = "Michael Renfrew and Worton, \{Bruce J\}",

year = "2024",

month = jul,

day = "12",

doi = "10.1007/978-3-031-65723-8\_20",

language = "English",

isbn = "978-3-031-65722-1",

series = "Developments in Statistical Modelling",

publisher = "Springer",

pages = "128--133",

editor = "J Einbeck and H Maeng and E Ogundimu and K Perrakis",

booktitle = "Developments in Statistical Modelling",

address = "United Kingdom",

}

A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data. / Renfrew, Michael; Worton, Bruce J.
Developments in Statistical Modelling: IWSM 2024. ed. / J Einbeck; H Maeng; E Ogundimu; K Perrakis. Springer, 2024. p. 128-133 (Developments in Statistical Modelling).

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data

AU - Renfrew, Michael

AU - Worton, Bruce J

PY - 2024/7/12

Y1 - 2024/7/12

N2 - The paper compares two advanced boosting techniques, Extreme Gradient Boosting (XGBoost) and Gaussian Process Boosting (GPBoost), for spatial logistic regression models using satellite data. It highlights the limitations of XGBoost in handling autocorrelated data typical in spatial statistics and demonstrates how GPBoost can incorporate a Gaussian process within a mixed-effects model to significantly reduce generalization error.

AB - The paper compares two advanced boosting techniques, Extreme Gradient Boosting (XGBoost) and Gaussian Process Boosting (GPBoost), for spatial logistic regression models using satellite data. It highlights the limitations of XGBoost in handling autocorrelated data typical in spatial statistics and demonstrates how GPBoost can incorporate a Gaussian process within a mixed-effects model to significantly reduce generalization error.

U2 - 10.1007/978-3-031-65723-8_20

DO - 10.1007/978-3-031-65723-8_20

M3 - Chapter (peer-reviewed)

SN - 978-3-031-65722-1

T3 - Developments in Statistical Modelling

SP - 128

EP - 133

BT - Developments in Statistical Modelling

A2 - Einbeck, J

A2 - Maeng, H

A2 - Ogundimu, E

A2 - Perrakis, K

PB - Springer

ER -

A comparison of extreme gradient and Gaussian process boosting for a spatial logistic regression on satellite data

Abstract

Publication series

Access to Document

Fingerprint

Cite this