plot.MclustBoostrap: Plot of bootstrap distributions for mixture model parameters

In mclust: Gaussian Mixture Modelling for Model-Based Clustering, Classification, and Density Estimation

Plot of bootstrap distributions for mixture model parameters

Description

Plots the bootstrap distribution of parameters as returned by the MclustBootstrap function.

Usage

## S3 method for class 'MclustBootstrap'
plot(x, what = c("pro", "mean", "var"), 
     show.parest = TRUE, show.confint = TRUE,
     hist.col = "grey", hist.border = "lightgrey", breaks = "Sturges", 
     col = "forestgreen", lwd = 2, lty = 3, 
     xlab = NULL, xlim = NULL, ylim = NULL, ...)

Arguments

x	Object returned by MclustBootstrap.
what	Character string specifying if mixing proportions ("pro"), component means ("mean") or component variances ("var") should be drawn.
show.parest	A logical specifying if the parameter estimate should be drawn as vertical line.
show.confint	A logical specifying if the resampling-based confidence interval should be drawn at the bottom of the graph. Confidence level can be provided as further argument conf.level; see summary.MclustBootstrap.
hist.col	The color to be used to fill the bars of the histograms.
hist.border	The color of the border around the bars of the histograms.
breaks	See the argument in function hist.
col, lwd, lty	The color, line width and line type to be used to represent the estimated parameters and confidence intervals.
xlab	Optional label for the horizontal axis.
xlim, ylim	A two-values vector of axis range for, respectively, horizontal and vertical axis.
...	Other graphics parameters.

Value

A plot for each variable/component of the selected parameters.

See Also

MclustBootstrap

Examples

data(diabetes)
X <- diabetes[,-1]
modClust <- Mclust(X, G = 3, modelNames = "VVV")
bootClust <- MclustBootstrap(modClust, nboot = 99)
par(mfrow = c(1,3), mar = c(4,2,2,0.5))
plot(bootClust, what = "pro")
par(mfrow = c(3,3), mar = c(4,2,2,0.5))
plot(bootClust, what = "mean")

mclust documentation built on Nov. 16, 2023, 5:10 p.m.