Measurement error-filtered machine learning in digital soil mapping
This paper presents a two-stage maximum likelihood framework to deal with measurement errors in digital soil mapping (DSM) when using a machine learning (ML) model. The framework is implemented with random forest and projection pursuit regression to illustrate two different areas of machine learning, i.e. ensemble learning with trees and feature-learning. In our proposed framework, a measurement error variance (MEV) is incorporated as a weight in the log-likelihood function so that measurements with a larger MEV receive less weight when a ML model is calibrated. We evaluate the performance of the error-filtered ML models with an error-filtered regression kriging model, in a comprehensive simulation study and in a real-world case study of Namibian data. From the results we show that prediction accuracy can be increased by using our proposed framework, especially when the MEVs are large and heterogeneous.
Main Authors: | , , , |
---|---|
Format: | Article/Letter to editor biblioteca |
Language: | English |
Subjects: | Digital soil mapping, E-BLUP, Machine learning, Maximum likelihood, Measurement error, REML, |
Online Access: | https://research.wur.nl/en/publications/measurement-error-filtered-machine-learning-in-digital-soil-mappi |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
dig-wur-nl-wurpubs-591525 |
---|---|
record_format |
koha |
spelling |
dig-wur-nl-wurpubs-5915252024-10-02 van der Westhuizen, Stephan Heuvelink, Gerard B.M. Hofmeyr, David P. Poggio, Laura Article/Letter to editor Spatial Statistics 47 (2022) ISSN: 2211-6753 Measurement error-filtered machine learning in digital soil mapping 2022 This paper presents a two-stage maximum likelihood framework to deal with measurement errors in digital soil mapping (DSM) when using a machine learning (ML) model. The framework is implemented with random forest and projection pursuit regression to illustrate two different areas of machine learning, i.e. ensemble learning with trees and feature-learning. In our proposed framework, a measurement error variance (MEV) is incorporated as a weight in the log-likelihood function so that measurements with a larger MEV receive less weight when a ML model is calibrated. We evaluate the performance of the error-filtered ML models with an error-filtered regression kriging model, in a comprehensive simulation study and in a real-world case study of Namibian data. From the results we show that prediction accuracy can be increased by using our proposed framework, especially when the MEVs are large and heterogeneous. en application/pdf https://research.wur.nl/en/publications/measurement-error-filtered-machine-learning-in-digital-soil-mappi 10.1016/j.spasta.2021.100572 https://edepot.wur.nl/561553 Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ Wageningen University & Research |
institution |
WUR NL |
collection |
DSpace |
country |
Países bajos |
countrycode |
NL |
component |
Bibliográfico |
access |
En linea |
databasecode |
dig-wur-nl |
tag |
biblioteca |
region |
Europa del Oeste |
libraryname |
WUR Library Netherlands |
language |
English |
topic |
Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML |
spellingShingle |
Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML van der Westhuizen, Stephan Heuvelink, Gerard B.M. Hofmeyr, David P. Poggio, Laura Measurement error-filtered machine learning in digital soil mapping |
description |
This paper presents a two-stage maximum likelihood framework to deal with measurement errors in digital soil mapping (DSM) when using a machine learning (ML) model. The framework is implemented with random forest and projection pursuit regression to illustrate two different areas of machine learning, i.e. ensemble learning with trees and feature-learning. In our proposed framework, a measurement error variance (MEV) is incorporated as a weight in the log-likelihood function so that measurements with a larger MEV receive less weight when a ML model is calibrated. We evaluate the performance of the error-filtered ML models with an error-filtered regression kriging model, in a comprehensive simulation study and in a real-world case study of Namibian data. From the results we show that prediction accuracy can be increased by using our proposed framework, especially when the MEVs are large and heterogeneous. |
format |
Article/Letter to editor |
topic_facet |
Digital soil mapping E-BLUP Machine learning Maximum likelihood Measurement error REML |
author |
van der Westhuizen, Stephan Heuvelink, Gerard B.M. Hofmeyr, David P. Poggio, Laura |
author_facet |
van der Westhuizen, Stephan Heuvelink, Gerard B.M. Hofmeyr, David P. Poggio, Laura |
author_sort |
van der Westhuizen, Stephan |
title |
Measurement error-filtered machine learning in digital soil mapping |
title_short |
Measurement error-filtered machine learning in digital soil mapping |
title_full |
Measurement error-filtered machine learning in digital soil mapping |
title_fullStr |
Measurement error-filtered machine learning in digital soil mapping |
title_full_unstemmed |
Measurement error-filtered machine learning in digital soil mapping |
title_sort |
measurement error-filtered machine learning in digital soil mapping |
url |
https://research.wur.nl/en/publications/measurement-error-filtered-machine-learning-in-digital-soil-mappi |
work_keys_str_mv |
AT vanderwesthuizenstephan measurementerrorfilteredmachinelearningindigitalsoilmapping AT heuvelinkgerardbm measurementerrorfilteredmachinelearningindigitalsoilmapping AT hofmeyrdavidp measurementerrorfilteredmachinelearningindigitalsoilmapping AT poggiolaura measurementerrorfilteredmachinelearningindigitalsoilmapping |
_version_ |
1813194681961414656 |