uhcdatasimulator: uhcdatasimulator

In aaarchmiller/uhcplots: Used-Habitat Calibration Plots (UHC Plots)

Description

uhcdatasimulator simulates the data used in Fieberg et al. (In Review)

Usage

uhcdatasimulator(nused, navail, betas, corx, ntemp, example)

Arguments

nused	The number of used locations in training/test data set
navail	The number of background locations in training/test data set
betas	The vector of length 2 for *true* probability of use
corx	The correlation between elevation and precipitation in training/test dataset. For missing predictor example only.
ntemp	A large number of available points.
example	The name of the example. Options include "missing predictor" or "non-linear".

Details

This is a function that creates a dataframe based on the example chosen from the manuscript (Fieberg et al. 2018). In the first example, ("missing predictor") the distribution of a species is related to elevation (x_1) and precipitation (x_2), where x_1 and x_2 are normally distributed with mean 0 and variance 4. We considered 3 different data-generating scenarios in which we varied the correlation of x_1,x_2 (corx) in the training and test data sets. In the second example ("non-linear"), the distribution of a species is non-linearly related to temperature (x_3 \sim N(0,4)).

Value

A dataframe of simulated data

See Also

Full archive of the data and code necessary to replicate the manuscript at http://doi.org/10.13020/D6T590.

Examples

# Simulate training or test data for the non-linear example
nonlinear.data <- uhcdatasimulator(nused = 100,
   navail = 10000,
   betas = c(2,-1),
   ntemp = 1000000,
   example = "non-linear")

# Simulate training or test data for the missing predictor example
   #   Where corr(x1,x2) = 0
missingpredictor.0.data <- uhcdatasimulator(nused = 100,
   navail = 10000,
   betas = c(0.5,-1),
   corx = 0,
   ntemp = 1000000,
   example = "missing predictor")

   #   Where corr(x1,x2) = -0.3
missingpredictor.N.data <- uhcdatasimulator(nused = 100,
   navail = 10000,
   betas = c(0.5,-1),
   corx = -0.3,
   ntemp = 1000000,
   example = "missing predictor")

   #   Where corr(x1,x2) = 0.3
missingpredictor.P.data <- uhcdatasimulator(nused = 100,
   navail = 10000,
   betas = c(0.5,-1),
   corx = 0.3,
   ntemp = 1000000,
   example = "missing predictor")

aaarchmiller/uhcplots documentation built on May 10, 2019, 2:05 a.m.