evaluate {modEvA} | R Documentation |
Evaluate a model based on the elements of a confusion matrix.
Description
This function evaluates the classification performance of a model based on the values of a confusion matrix obtained at a particular threshold.
Usage
evaluate(a, b, c, d, N = NULL, measure = "CCR")
Arguments
a |
number of correctly predicted presences |
b |
number of absences incorrectly predicted as presences |
c |
number of presences incorrectly predicted as absences |
d |
number of correctly predicted absences |
N |
total number of cases. If NULL (the default), it is calculated automatically by adding up a, b, c and d.) |
measure |
a character vector of length 1 indicating the evaluation measure to use. Type |
Details
A number of measures can be used to evaluate continuous model predictions against observed binary occurrence data (Fielding & Bell 1997; Liu et al. 2011; Barbosa et al. 2013; Leroy et al. 2018). The 'evaluate' function can calculate a few threshold-based classification measures from the values of a confusion matrix obtained at a particular threshold. The 'evaluate' function is used internally by threshMeasures
. It can also be accessed directly by the user, but it is usually more practical to use 'threshMeasures', which calculates the confusion matrix automatically.
Value
The value of the specified evaluation measure.
Note
Some measures (e.g. NMI, odds ratio) don't work with zeros in (some parts of) the confusion matrix. Also, TSS and NMI are not symmetrical, i.e. "obs" vs "pred" is different from "pred" vs "obs".
Author(s)
A. Marcia Barbosa
References
Barbosa A.M., Real R., Munoz A.R. & Brown J.A. (2013) New measures for assessing model equilibrium and prediction mismatch in species distribution models. Diversity and Distributions, 19: 1333-1338
Fielding A.H. & Bell J.F. (1997) A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation 24: 38-49
Leroy B., Delsol R., Hugueny B., Meynard C.M., Barhoumi C., Barbet-Massin M. & Bellard C. (2018) Without quality presence-absence data, discrimination metrics such as TSS can be misleading measures of model performance. Journal of Biogeography 45(9):1994-2002
Liu C., White M., & Newell G. (2011) Measuring and comparing the accuracy of species distribution models with presence-absence data. Ecography, 34, 232-243.
See Also
Examples
evaluate(23, 44, 21, 34)
evaluate(23, 44, 21, 34, measure = "TSS")