make_kernel_dist: Calculate 1D, 2D, or 3D kernel smooth distribution

In simlandr: Simulation-Based Landscape Construction for Dynamical Systems

Calculate 1D, 2D, or 3D kernel smooth distribution

Description

Calculate 1D, 2D, or 3D kernel smooth distribution

Usage

make_kernel_dist(
  output,
  var_names,
  lims,
  kde_fun,
  n,
  h,
  adjust,
  weight_var = NULL
)

Arguments

output	A matrix of simulation output, or a mcmc, mcmc.list object (see coda::mcmc()).
var_names	The names of the target variables.
lims	The limits of the range for the density estimator as c(xl, xu) for 2D landscapes, c(xl, xu, yl, yu) for 3D landscapes, c(xl, xu, yl, yu, zl, zu) for 4D landscapes. If missing, the range of the data extended by 10% for both sides will be used. For landscapes based on multiple simulations, the largest range of all simulations (which means the lowest lower limit and the highest upper limit) will be used by default.
kde_fun	Which kernel estimator to use? Choices: "ks" ks::kde() (default; faster and using less memory); "base" base::density() (only for 2D landscapes); "MASS" MASS::kde2d() (only for 3D landscapes).
n	The number of equally spaced points in each axis, at which the density is to be estimated.
h	A number, or possibly a vector for 3D and 4D landscapes, specifying the smoothing bandwidth to be used. If missing, the default value of the kernel estimator will be used (but bw = "SJ" for base::density()). Note that the definition of bandwidth might be different for different kernel estimators. For landscapes based on multiple simulations, the largest h of all simulations will be used by default.
adjust	The multiplier to the bandwidth. The bandwidth used is actually adjust * h. This makes it easy to specify values like "half the default" bandwidth.
weight_var	The name of the weight variable, in case the weight of each observation is different. This may be useful when a weighted MC (e.g., importance sampling) is used. Only effective for kde_fun = "ks".

Value

A list of the smooth distribution.

simlandr documentation built on April 4, 2025, 1:14 a.m.