BASiCS_DiagPlot: Create diagnostic plots of MCMC parameters

In BASiCS: Bayesian Analysis of Single-Cell Sequencing data

Description

Plot parameter values and effective sample size. See effectiveSize for more details on this diagnostic measure.

Usage

BASiCS_DiagPlot(
  object,
  Parameter = "mu",
  Measure = c("ess", "geweke.diag"),
  x = NULL,
  y = NULL,
  LogX = isTRUE(x %in% c("mu", "delta")),
  LogY = isTRUE(y %in% c("mu", "delta")),
  Smooth = TRUE,
  na.rm = TRUE
)

BASiCS_diagPlot(...)

Arguments

object	an object of class BASiCS_Summary
Parameter	Optional name of a chain parameter to restrict the histogram; if not supplied, all parameters will be assessed. Possible values: 'mu', 'delta', 'phi', 's', 'nu', 'theta', 'beta', 'sigma2' and 'epsilon'. Default Parameter = 'mu'
Measure	Character scalar specifying the diagnostic measure to plot. Current options are effective sample size and the Geweke diagnostic criterion.
x, y	Optional MCMC parameter values to be plotted on the x or y axis, respectively. If neither is supplied, Parameter will be plotted on the x axis and effective sample size will be plotted on the y axis as a density plot.
LogX, LogY	A logical value indicating whether to use a log10 transformation for the x or y axis, respectively.
Smooth	A logical value indicating whether to use smoothing (specifically hexagonal binning using geom_hex).
na.rm	Logical value indicating whether NA values should be removed before calculating effective sample size.
...	Unused.

Value

A ggplot object.

Author(s)

Alan O'Callaghan

See Also

BASiCS_Chain

Examples

# Built-in example chain
data(ChainSC)

# Point estimates versus effective sample size
BASiCS_DiagPlot(ChainSC, Parameter = "mu")
# Effective sample size as colour, mu as x, delta as y.
BASiCS_DiagPlot(ChainSC, x = "mu", y = "delta")

BASiCS documentation built on April 16, 2021, 6 p.m.