Assessing the geographic specificity of pH prediction by classification and regression trees.

Soil pH effects a wide range of critical biogeochemical processes that dictate plant growth and diversity. Previous literature has established the capacity of classification and regression trees (CARTs) to predict soil pH, but limitations of CARTs in this context have not been fully explored. The cu...

Full description

Bibliographic Details
Main Authors: Jacob Egelberg, Nina Pena, Rachel Rivera, Christina Andruk
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2021-01-01
Series:PLoS ONE
Online Access:https://doi.org/10.1371/journal.pone.0255119
id doaj-dc5a9484cb8f4f3b928cfad97cd70d4b
record_format Article
spelling doaj-dc5a9484cb8f4f3b928cfad97cd70d4b2021-08-17T04:31:13ZengPublic Library of Science (PLoS)PLoS ONE1932-62032021-01-01168e025511910.1371/journal.pone.0255119Assessing the geographic specificity of pH prediction by classification and regression trees.Jacob EgelbergNina PenaRachel RiveraChristina AndrukSoil pH effects a wide range of critical biogeochemical processes that dictate plant growth and diversity. Previous literature has established the capacity of classification and regression trees (CARTs) to predict soil pH, but limitations of CARTs in this context have not been fully explored. The current study collected soil pH, climatic, and topographic data from 100 locations across New York's Temperate Deciduous Forests (in the United States of America) to investigate the extrapolative capacity of a previously developed CART model as compared to novel CART and random forest (RF) models. Results showed that the previously developed CART underperformed in terms of predictive accuracy (RRMSE = 14.52%) when compared to a novel tree (RRMSE = 9.33%), and that a novel random forest outperformed both models (RRMSE = 8.88%), though its predictions did not differ significantly from the novel tree (p = 0.26). The most important predictors for model construction were climatic factors. These findings confirm existing reports that CART models are constrained by the spatial autocorrelation of geographic data and encourage the restricted application of relevant machine learning models to regions from which training data was collected. They also contradict previous literature implying that random forests should meaningfully boost the predictive accuracy of CARTs in the context of soil pH.https://doi.org/10.1371/journal.pone.0255119
collection DOAJ
language English
format Article
sources DOAJ
author Jacob Egelberg
Nina Pena
Rachel Rivera
Christina Andruk
spellingShingle Jacob Egelberg
Nina Pena
Rachel Rivera
Christina Andruk
Assessing the geographic specificity of pH prediction by classification and regression trees.
PLoS ONE
author_facet Jacob Egelberg
Nina Pena
Rachel Rivera
Christina Andruk
author_sort Jacob Egelberg
title Assessing the geographic specificity of pH prediction by classification and regression trees.
title_short Assessing the geographic specificity of pH prediction by classification and regression trees.
title_full Assessing the geographic specificity of pH prediction by classification and regression trees.
title_fullStr Assessing the geographic specificity of pH prediction by classification and regression trees.
title_full_unstemmed Assessing the geographic specificity of pH prediction by classification and regression trees.
title_sort assessing the geographic specificity of ph prediction by classification and regression trees.
publisher Public Library of Science (PLoS)
series PLoS ONE
issn 1932-6203
publishDate 2021-01-01
description Soil pH effects a wide range of critical biogeochemical processes that dictate plant growth and diversity. Previous literature has established the capacity of classification and regression trees (CARTs) to predict soil pH, but limitations of CARTs in this context have not been fully explored. The current study collected soil pH, climatic, and topographic data from 100 locations across New York's Temperate Deciduous Forests (in the United States of America) to investigate the extrapolative capacity of a previously developed CART model as compared to novel CART and random forest (RF) models. Results showed that the previously developed CART underperformed in terms of predictive accuracy (RRMSE = 14.52%) when compared to a novel tree (RRMSE = 9.33%), and that a novel random forest outperformed both models (RRMSE = 8.88%), though its predictions did not differ significantly from the novel tree (p = 0.26). The most important predictors for model construction were climatic factors. These findings confirm existing reports that CART models are constrained by the spatial autocorrelation of geographic data and encourage the restricted application of relevant machine learning models to regions from which training data was collected. They also contradict previous literature implying that random forests should meaningfully boost the predictive accuracy of CARTs in the context of soil pH.
url https://doi.org/10.1371/journal.pone.0255119
work_keys_str_mv AT jacobegelberg assessingthegeographicspecificityofphpredictionbyclassificationandregressiontrees
AT ninapena assessingthegeographicspecificityofphpredictionbyclassificationandregressiontrees
AT rachelrivera assessingthegeographicspecificityofphpredictionbyclassificationandregressiontrees
AT christinaandruk assessingthegeographicspecificityofphpredictionbyclassificationandregressiontrees
_version_ 1721205589605875712