1997
DOI: 10.21236/ada325255
|View full text |Cite
|
Sign up to set email alerts
|

An Alternative to Correspondence Analysis Using Hellinger Distance.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
151
0
2

Year Published

2005
2005
2021
2021

Publication Types

Select...
10

Relationship

0
10

Authors

Journals

citations
Cited by 135 publications
(153 citation statements)
references
References 0 publications
0
151
0
2
Order By: Relevance
“…A multivariate regression tree (MRT) followed the approach described by Borcard et al (2011) and Lindfield et al (2016) using the R package mvpart (De'Ath, 2007) with MRTs primarily acting as predictive (vs. explanatory) models. Prior to MRT generation, relative abundance data were first Hellinger-transformed, which is an approach well-suited for species abundance datasets, granting lower weights to rare species (Legendre and Gallagher, 2001) and multiple zero counts (Rao, 1995). Optimal tree size was generated from 100 model runs, with the model selection output based on the highest cross-validated predictive accuracy.…”
Section: Habitat and Environmental Linkagesmentioning
confidence: 99%
“…A multivariate regression tree (MRT) followed the approach described by Borcard et al (2011) and Lindfield et al (2016) using the R package mvpart (De'Ath, 2007) with MRTs primarily acting as predictive (vs. explanatory) models. Prior to MRT generation, relative abundance data were first Hellinger-transformed, which is an approach well-suited for species abundance datasets, granting lower weights to rare species (Legendre and Gallagher, 2001) and multiple zero counts (Rao, 1995). Optimal tree size was generated from 100 model runs, with the model selection output based on the highest cross-validated predictive accuracy.…”
Section: Habitat and Environmental Linkagesmentioning
confidence: 99%
“…Otherwise, the difference in species composition between the years would have confounded further multivariate analyses. To prevent biases introduced by a large number of zero cover occurrences or by a small number of very high cover values, the cover values in the species data matrix were transformed using the Hellinger distance [30], which is the square root of the row totals divided by the row mean values. This distance measure showed good performance for abundance data throughout different multivariate ordination methods, especially for those relying on a linear gradient structure [31].…”
Section: Vegetation Samplingmentioning
confidence: 99%
“…In order to reduce the influence of less abundant species, all the species with less than 1% accumulated occurrence were excluded from the abundance matrix (SEDBERRY;CARTE, 1993;FEYRER;HEALEY, 2003). The reduced matrix data were then submitted to an F max test and transformed into Ln (x + 1) form, before further transformation through the Hellinger´s distance (RAO, 1995;LEGENDRE;LEGENDRE, 1998).…”
Section: Data Collectionmentioning
confidence: 99%