Partial least squares regression as an alternative to current regression methods used in ecology

Abstract

This paper briefly presents the aims, requirements and results of partial least squares regression analysis (PLSR), and its potential utility in ecological studies. This statistical technique is particularly well suited to analyzing a large array of related predictor variables (i.e. not truly independent), with a sample size not large enough compared to the number of independent variables, and in cases in which an attempt is made to approach complex phenomena or syndromes that must be defined as a combination of several variables obtained independently. A simulation experiment is carried out to compare this technique with multiple regression (MR) and with a combination of principal component analysis and multiple regression (PCA+MR), varying the number of predictor variables and sample sizes. PLSR models explained a similar amount of variance to those results obtained by MR and PCA+MR. However, PLSR was more reliable than other techniques when identifying relevant variables and their magnitudes of influence, especially in cases of small sample size and low tolerance. Finally, we present one example of PLSR to illustrate its application and interpretation in ecology.

Long-term climate-associated dynamics of a tropical seagrass meadow: implications for the future

“…Our study had 16 observations (years) and 30 variables, and many of the variables were collinear. This technique has commonly been used to analyse a range of ecological data sets (Amand et al 2004, Larocque et al 2006, Carrascal et al 2009). …”

“…PLS regression is an extension of multiple regression analysis in which the effects of linear combinations of several predictors on a response variable (or multiple response variables) are analysed. PLS regression is particularly suited to incidences when the matrix of predictors has more variables than observations, and when there is multicollinearity among variables (Carrascal et al 2009). Our study had 16 observations (years) and 30 variables, and many of the variables were collinear.…”

Habitat Configuration Alters Herbivory across the Tropical Seascape

“…PLS was used to investigate which of the observed seagrass meadow variables correlated most with epicover. PLS regression is particularly suited to incidences when there is multi co-linearity among variables (Carrascal et al, 2009). The study had variables that were co-linear.…”

“…In order to examine the potential influence of herbivory on attached epicover a Partial Least Squares Regression (PLS) model was developed in Minitab (version 17) (Carrascal et al, 2009;Haapkylä et al, 2011). PLS was used to investigate which of the observed seagrass meadow variables correlated most with epicover.…”

Factors affecting wild rabbit production in extensive breeding enclosures: how can we optimise efforts?

“…PLSR is a useful regression calibration technique when the number of predictor variables is similar to or higher than the number of observations and/or the predictors are highly correlated (Carrascal et al, 2009), and it reduces the exploratory variables into a few components that have maximum covariance with the dependent variable. A PLSR should therefore be used to deal with the structure of our data with 23 cases and 10 exploratory variables.…”