For building and evaluating species distribution models, the **porportion of presences** of the species and the **balance between the number of presences and absences** may be issues to take into account (e.g. Jiménez-Valverde & Lobo 2006, Barbosa et al. 2013). The *prevalence* and the *evenness* functions, included in the modEvA package (Barbosa et al. 2014), can calculate these measures:

prevalence <- function (obs, event = 1) {
# version 1.0
# calculates the prevalence (proportion of occurrences) of a value (event) in a vector
# obs: a vector of binary observations (e.g. 1 vs. 0, male vs. female, disease vs. no disease, etc.)
# event: the value whose prevalence we want to calculate (e.g. 1, "present", etc.)
sum(obs == event) / length(obs)
} # end prevalence function

evenness <- function (obs) {
# version 1.3 (18 June 2013)
# calculates the evenness (equilibrium) of cases in a binary vector; result ranges from 0 when all values are the same, to 1 when there are the same number of cases with each value
# obs: a vector of binary observations (e.g. 1 vs. 0, male vs. female, disease vs. no disease, etc.)
values <- unique(obs)
nvalues <- length(values)
if (!(nvalues %in% c(1, 2))) stop("Input vector includes ", nvalues, " different values; 'evenness' is only implemented for binary observations (with 1 or 2 different values).")
proportion <- (sum(obs == values[1])) / length(obs)
if (proportion > 0.5) balance <- 1 - proportion else balance <- proportion
return(2 * balance) # so evenness ranges between 0 and 1
} # end evenness function

Let’s exemplify with 3 sample binary vectors *x*, *y *and* z*:

(x <- rep(c(0, 1), each = 5))
(y <- c(rep(0, 3), rep(1, 7)))
(z <- c(rep(0, 7), rep(1, 3)))
prevalence(x)
evenness(x)
prevalence(y)
evenness(y)
prevalence(z)
evenness(z)

[presented with Pretty R]

**References**

Barbosa A.M., Brown J.A. & Real R. (2014) modEvA – an R package for model evaluation and analysis. R package, version 0.1.

Barbosa A.M., Real R., Muñoz 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

Jiménez-Valverde A. & Lobo J.M. (2006) The ghost of unbalanced species distribution data in geographical model predictions. *Diversity and Distributions*, **12**: 521–524.

### Like this:

Like Loading...

*Related*