DOI: 10.1007/978-3-642-05029-9_13
关键词: Linear model 、 Computer science 、 Species distribution 、 Regression 、 Environmental niche modelling 、 Collinearity 、 Dimensional reduction 、 Spatial analysis 、 Limit (mathematics) 、 Econometrics
摘要: Species distribution models have become a commonplace exercise over the last 10 years, however, analyses vary due to different traditions, aims of applications and statistical backgrounds. In this chapter, I lay out what consider be most crucial steps in species analysis: data pre-processing visualisation, dimensional reduction (including collinearity), model formulation, simplification, type, assessment performance (incl. spatial autocorrelation) interpretation. For each step, relevant considerations are discussed, mainly illustrated with Generalised Linear Models Boosted Regression Trees as two contrasting methods. second section, draw attention three challenging problems modelling: identifying (and incorporating into model) factors that limit range; separating fundamental, realised potential niche; niche evolution.