R: Plots for Mixture-Based Density Estimate

plot.densityMclust {mclust}

R Documentation

Plots for Mixture-Based Density Estimate

Description

Plotting methods for an object of class 'mclustDensity'. Available graphs are plot of BIC values and density for univariate and bivariate data. For higher data dimensionality a scatterplot matrix of pairwise densities is drawn.

Usage

## S3 method for class 'densityMclust'
plot(x, data = NULL, what = c("BIC", "density", "diagnostic"), ...)

plotDensityMclust1(x, data = NULL, hist.col = "lightgrey", 
                   hist.border = "white",  breaks = "Sturges", ...)

plotDensityMclust2(x, data = NULL, nlevels = 11, levels = NULL, col = grey(0.6),
                   points.pch = 1, points.col = 1, points.cex = 0.8, ...)

plotDensityMclustd(x, data = NULL, nlevels = 11, levels = NULL, col = grey(0.6), 
                   points.pch = 1, points.col = 1, points.cex = 0.8,
                   gap = 0.2, ...)

Arguments

`x`	An object of class `'mclustDensity'` obtained from a call to `densityMclust` function.
`data`	Optional data points.
`what`	The type of graph requested: `"density"` = a plot of estimated density; if `data` is also provided the density is plotted over data points (see Details section). `"BIC"` = a plot of BIC values for the estimated models versus the number of components. `"diagnostic"` = diagnostic plots (only available for the one-dimensional case, see `densityMclust.diagnostic`)
`hist.col`	The color to be used to fill the bars of the histogram.
`hist.border`	The color of the border around the bars of the histogram.
`breaks`	See the argument in function `hist`.
`points.pch, points.col, points.cex`	The character symbols, colors, and magnification to be used for plotting `data` points.
`nlevels`	An integer, the number of levels to be used in plotting contour densities.
`levels`	A vector of density levels at which to draw the contour lines.
`col`	Color to be used for drawing the contour lines, the perspective plot, or the image density. In the latter case can be also a vector of color values.
`gap`	Distance between subplots, in margin lines, for the matrix of pairwise scatterplots.
`...`	Additional arguments.

Details

The function plot.densityMclust allows to obtain the plot of estimated density or the graph of BIC values for evaluated models.

If what = "density" the produced plot dependes on the dimensionality of the data.

For one-dimensional data a call with no data provided produces a plot of the estimated density over a sensible range of values. If data is provided the density is over-plotted on a histogram for the observed data.

For two-dimensional data further arguments available are those accepted by the surfacePlot function. In particular, the density can be represented through "contour", "image", and "persp" type of graph.

For higher dimensionality a scatterplot matrix of pairwise densities is drawn.

References

C. Fraley and A. E. Raftery (2002). Model-based clustering, discriminant analysis, and density estimation. Journal of the American Statistical Association 97:611:631.

C. Fraley, A. E. Raftery, T. B. Murphy and L. Scrucca (2012). mclust Version 4 for R: Normal Mixture Modeling for Model-Based Clustering, Classification, and Density Estimation. Technical Report No. 597, Department of Statistics, University of Washington.

Author(s)

Luca Scrucca

Examples

dens = densityMclust(faithful$waiting)
plot(dens, what = "density")
plot(dens, what = "density", data = faithful$waiting)

dens = densityMclust(faithful)
plot(dens, what = "density")
plot(dens, what = "density", type = "image", col = "steelblue")
plot(dens, what = "density", type = "persp", col = adjustcolor("steelblue", alpha.f = 0.5))

x = iris[,1:4]
dens = densityMclust(x)
plot(dens, what = "density", nlevels = 7)
## Not run: 
  plot(dens, x, what = "density", drawlabels = FALSE,
       levels = quantile(dens$density, probs = c(0.05, 0.25, 0.5, 0.75, 0.95)))
  plot(dens, what = "density", type = "image", col = "steelblue")
  plot(dens, what = "density", type = "persp", border = adjustcolor(grey(0.1), alpha.f = 0.5))

## End(Not run)

[Package mclust version 5.3 Index]