pROC applies partial ROC tests to continuous niche models.
pROC( continuous_mod, test_data, n_iter = 1000, E_percent = 5, boost_percent = 50, parallel = FALSE, ncores = 4, rseed = FALSE )
| continuous_mod | a RasterLayer or a numeric vector of the ecological niche model to be evaluated. If a numeric vector is provided it should contain the values of the predicted suitability. |
|---|---|
| test_data | A numerical matrix, data.frame, or a numeric vector. If it is data.frame or matrix it should contain coordinates of the occurrences used to test the ecological niche model to be evaluated; columns must be: longitude and latitude. If numeric vector it should contain the values of the predicted suitability. |
| n_iter | (numeric) number of bootstrap iterations to be performed; default = 1000. |
| E_percent | (numeric) value from 0 to 100 that will be used as a threshold (E); default = 5. |
| boost_percent | (numeric) value from 0 to 100 representing the percent of testing data to be used for performing the bootstrap process for calculating the partial ROC; default = 50. |
| parallel | Logical to specify if the computation will be done in parallel. default=TRUE. |
| ncores | Numeric; the number of cores to be used for parallelization. |
| rseed | Logical. Whether or not to set a random seed. Default FALSE. |
A data.frame containing the AUC values and AUC ratios calculated for each iteration.
Partial ROC is calculated following Peterson et al. (2008; http://dx.doi.org/10.1016/j.ecolmodel.2007.11.008). This function is a modification of the PartialROC funcion, available at https://github.com/narayanibarve/ENMGadgets.
Peterson, A.T. et al. (2008) Rethinking receiver operating characteristic analysis applications in ecological niche modeling. Ecol. Modell., 213, 63–72.
# Load a continuous model conti_model <- raster::raster(system.file("extdata", "ambystoma_model.tif", package="ntbox")) # Read validation (test) data test_data <- read.csv(system.file("extdata", "ambystoma_validation.csv", package = "ntbox")) # Filter only presences as the Partial ROC only needs occurrence data test_data <- dplyr::filter(test_data, presence_absence==1) test_data <- test_data[,c("longitude","latitude")] partial_roc <- pROC(continuous_mod=conti_model, test_data = test_data, n_iter=1000,E_percent=5, boost_percent=50, parallel=FALSE)