caretSDM

Luíz Fernando Esser

caretSDM

caretSDM is a under development R package that uses the powerful caret package as the main engine to obtain Species Distribution Models. As caret is a packaged turned to build machine learning models, caretSDM has a strong focus on this approach.

Installation

You can install the development version of caretSDM from GitHub with:

install.packages("devtools")
devtools::install_github("luizesser/caretSDM")

The package is also available on CRAN. Users are able to install it using the following code:

install.packages("caretSDM")

You need help?

caretSDM is vastly documented and has included some objects that can guide your data management. If some of your data or code seem to be wrong, try to take a look at those objects or the articles in the website:

Objects

• bioc Bioclimatic variables for current scenario in stars class.

• rivs Hydrological variables for current scenario in sf class.

• occ Araucaria angustifolia occurrence data as a dataframe.

• salm Salminus brasiliensis occurrence data as a dataframe.

• parana Shapefile to use in sdm_area in Simple Feature class.

• scen Bioclimatic variables for future scenarios in stars class.

• scen_rs Bioclimatic variables for invasive assessments vignette.

• algorithms Dataframe with characteristics from every algorithm available in caretSDM.

Articles

• 1. Concatenate functions in caretSDM This vignette shows how to build compact scripts, which is very useful to run your first tests.

• 2. Adding New Algorithms to caretSDM Do not found your ideal algorithm already implemented? Here we show how to implement any custom algorithm in our package.

• 3. caretSDM Workflow for Species Distribution Modeling This is the main vignette for terrestrial species modeling, where we model the tree species Araucaria angustifolia.

• 4. Modeling Species Distributions in Continental Water Bodies This is the main vignette for continental aquatic species modeling, where we model the fish species Salminus brasiliensis.

• 5. Projecting Non-native Distribution using SDMs Here we demonstrate how to make invasiveness assessments using the package.

• 6. Modeling Rare Species using Ensemble of Small Models Want to model rare species? This vignette showcases how easy it is to apply SDMs to rare species with low number of records.

• 7. Applying MaxEnt in caretSDM You are a MaxEnt-only type of modeler? Then this vignette is for you. Here we show how to obtain SDMs using MaxEnt with automatic feature selection.

• 8. Benchmarking SDM Package Performance in R Want to know how this package performs in comparison with other popular packages? Here we address this question through benchmarking.

• 9. Comparing Pseudoabsence Methods in Species Distribution Modelling To build a script to compare different approaches in SDMs you can use this vignette as a starting-point.

• 10. Ablation Analysis of Hyperparameter Tuning Length in caretSDM Unveils the impact of searching for optimized hyperparameters in the final models.