Skip to contents

Create a spreadsheet containing all SDM performance metrics

make_evaluation_csv.Rd

Loads existing performance metrics for component models, and calculates Area under the Curve (AUC) for the ensemble model 2 ways. This function also has the option to test the model on an external dataset using test_ens. The final product is a saved csv file containing all of the performance metrics.

Usage

make_evaluation_csv(spp_list, test_ens = T, yr_min, yr_max)

Arguments

spp_list

the data.frame containing species names and alternative names for test_ens. Must contain the column Name, which matches the names of the species folders to help pull correct data.

test_ens

a TRUE/FALSE indicating whether or not to test the ensemble model on an external dataset

yr_min

start of year range to specify which predictions to use. Only used if test_ens == TRUE

yr_max

end of year range to specify which predictions to use. Only used if test_ens == TRUE

Value

nothing is returned. The resulting CSV file called 'species_evaluation_metrics.csv' is saved in the working directory.

Details

The resulting CSV file contains the following columns:

Common.Name

name of species

Managing.Body

name of group responsible for species management, if exists

Feeding.Guild

assigned feeding guild

Habitat.Guild

assigned habitat guild

N.PRESENCE, N.ABSENCE

number of presences/absences in the data set used to build the model

BRT, GAM, MAXENT, RF, SDMTMB

AUC for each component model

BRT.WT, GAM.WT, MAXENT.WT, RF.WT, SDMTMB.WT

weights for each component model in the final ensemble

ENS.AUC

Ensemble model AUC based on weighted sum of CV predictions from each component model

WAVG.ENS.AUC

Ensemble model AUC calculated as a weighted average of AUCs, using the corresponding weights

AUC.yr_min.yr_max

Ensemble model AUC calculated on the external data, if test_ens == TRUE. Column name will reflect the timeseries used