Abstract
Species distribution modeling often involves high‐dimensional environmental data. Large amounts of data and multicollinearity among covariates impose challenges to statistical models in variable selection for reliable inferences of the effects of environmental factors on the spatial distribution of species. Few studies have evaluated and compared the performance of multiple machine learning (ML) models in handling multicollinearity. Here, we assessed the effectiveness of removal of correlated covariates and regularization to cope with multicollinearity in ML models for habitat suitability. Three machine learning algorithms maximum entropy (MaxEnt), random forests (RFs), and support vector machines (SVMs) were applied to the original data (OD) of 27 landscape variables, reduced data (RD) with 14 highly correlated covariates being removed, and 15 principal components (PC) of the OD accounting for 90% of the original variability. The performance of the three ML models was measured with the area under the curve and continuous Boyce index. We collected 663 nonduplicated presence locations of Eastern wild turkeys (Meleagris gallopavo silvestris) across the state of Mississippi, United States. Of the total locations, 453 locations separated by a distance of ≥2 km were used to train the three ML algorithms on the OD, RD, and PC data, respectively. The remaining 210 locations were used to validate the trained ML models to measure ML performance. Three ML models had excellent performance on the RD and PC data. MaxEnt and SVMs had good performance on the OD data, indicating the adequacy of regularization of the default setting for multicollinearity. Weak learning of RFs through bagging appeared to alleviate multicollinearity and resulted in excellent performance on the OD data. Regularization of ML algorithms may help exploratory studies of the effects of environmental factors on the spatial distribution and habitat suitability of wildlife.
Highlights
Studies of the spatiotemporal distribution of resources that sup‐ port organisms are indispensable for understanding the dynamics of animal populations, including avian populations, across space and time (Fuller, 2012)
We evaluated the predictive accuracy of Ecological niche factor analysis (ENFA), random forests (RFs), maximum entropy (MaxEnt), and support vector machines (SVMs) predictions using the same test data (210 nonduplicated presence locations) with the continuous Boyce index (CBI; Boyce, Vernier, Nielsen, & Schmiegelow, 2002; Hirzel, Lay, Helfer, Randin, & Guisan, 2006)
This study assessed the effectiveness of two different methods of correlation removal and principal component approaches to address multicollinearity on the predictive performance of Maximum en‐ tropy (MaxEnt), random forests (RFs), and support vector machines (SVMs) for habitat suitability modeling
Summary
Studies of the spatiotemporal distribution of resources that sup‐ port organisms are indispensable for understanding the dynamics of animal populations, including avian populations, across space and time (Fuller, 2012). Habitat suitability mapping often uses a large number of land‐ scape variables (e.g., 10 or more variables) to predict habitat suit‐ ability Many of those landscape variables are highly correlated to one another, leading to multicollinearity in habitat and resource selection models (Aebischer, Robertson, & Kenward, 1993; Cutler et al, 2007). Drake et al (2006) demonstrated that unprocessed data (their model 1) and orthogonal transformation (method 2) performed and better than correlation removal (method 3) in SVMs. Random forests may alleviate multicollinear‐ ity with a randomized subset of explanatory variables when grow‐ ing each tree branch (Cutler et al, 2007). We first developed statewide habitat suitability maps with a large sample size of presence data (e.g., 600–700 presence locations) using MaxEnt, RFs, and SVMs. Second, we compared predictive performances of MaxEnt, RFs, and SVMs between correlation removal and principal component ap‐ proaches to multicollinearity. Ecological studies have not exploited extensively the excellent performances of SVMs in pattern identifi‐ cation and recognition and the capacity to analyze large amounts of data and complex relationships (Huettmann et al, 2018)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
