Detecting ecological breakpoints: a new tool for piecewise regression

Publication Type  Journal Article
Year of Publication  2011
Authors  Alessandro Ferrarini
Journal Title  Computational Ecology and Software
Volume  1
Issue  2
Pages  121-124
Journal Date  06/2011
Short Title  Detecting ecological breakpoints: a new tool for piecewise regression
Publisher  International Academy of Ecology and Environmental Sciences
ISSN Number  2220-721X
Key Words  ecological breakpoints; piecewise regression; rolling regression; moving window regression; GIS.
Abstract  

Simple linear regression tries to determine a linear relationship between a given variable X (predictor) and a dependent variable Y. Since most of the environmental problems involve complex relationships, X-Y relationship is often better modeled through a regression where, instead of fitting a single straight line to the data, the algorithm allows the fitting to bend. Piecewise regressions just do it, since they allow emphasize local, instead of global, rules connecting predictor and dependent variables. In this work, a tool called RolReg is proposed as an implementation of Krummel's method to detect breakpoints in regression models. RolReg, which is freely available upon request from the author, could useful to detect proper breakpoints in ecological laws.

URL  http://www.iaees.org/publications/journals/ces/articles/2011-1(2)/Detecting-ecological-breakpoints.pdf
Export  Tagged XML BibTex