mkde (version 0.4)

mkde2Dgrid: Movement-based kernel density estimate (MKDE) in 2D using Rcpp

Description

Provides a function for 2-dimensional MKDEs.

Usage

mkde2Dgrid(mkde.obj, move.dat, t.step, d.thresh)

Value

An array whose elements are the estimated utilization probabilities for each cell.

Arguments

mkde.obj

A 2D or 2.5D MKDE object

move.dat

A move data object

t.step

An integration time step

d.thresh

A kernel density threshold

Author

Jeff A. Tracey, PhD
USGS Western Ecological Research Center, San Diego Field Station

James Sheppard, PhD
San Diego Zoo Institute for Conservation Research

Robert Sinkovits, PhD
San Diego Supercomputer Center

Glenn Lockwood, PhD
San Diego Supercomputer Center

Jun Zhu, PhD
University of Wisconsin-Madison

Details

This is lower-level function that call the C++ function. for estimating the movement-based density in 2D. In practice, users should call initializeDensity.
The argument d.thresh is a univariate probability density beyond which the kernel contribution to the overall MKDE is assumed to be negligible. Usually this is set at a very small value and is used to prevent calculations from being performed in cells to which the kernel makes a negligible contribution.

Examples

data(panda)

# Find min/max coordinates and add buffer
xmax = max(panda$x) + 100
xmin = min(panda$x) - 100
ymax = max(panda$y) + 100
ymin = min(panda$y) - 100

# Calculate grid dimensions
xrange <- xmax - xmin
yrange <- ymax - ymin
cell.sz = 30
nx <- as.integer(xrange/cell.sz)
ny <- as.integer(yrange/cell.sz)

mv.dat <- initializeMovementData(panda$time, panda$x, panda$y, t.max=185.0, sig2obs=25.0)
if (all(is.na(mv.dat$sig2xy))) {
  mv.dat <- estVarMKDE(mv.dat)
}

mkde.obj <- initializeMKDE2D(xmin, cell.sz, nx, ymin, cell.sz, ny)
res <- mkde2Dgrid(mkde.obj, mv.dat, 10.0, 1e-20)