plot.fabMix.object: Plot function

In fabMix: Overfitting Bayesian Mixtures of Factor Analyzers with Parsimonious Covariance and Unknown Number of Components

Plot function

Description

This function plots fabMix function.

Usage

## S3 method for class 'fabMix.object'
plot(x, what, variableSubset, class_mfrow, sig_correlation, confidence, ...)

Arguments

x	An object of class fabMix.object, which is returned by the fabMix function.
what	One of the "BIC", "classification_matplot", "classification_pairs", "correlation", "factor_loadings". The plot will display the BIC values per model and number of factors (along with the most probable number of clusters as text), a matplot per cluster for the selected model, scatterplots pairs, the estimated correlation matrix per cluster, and the MAP estimate of factor loadings, respectively.
variableSubset	An optional subset of the variables. By default, all variables are selected.
class_mfrow	An optional integer vector of length 2, that will be used to set the mfrow for "classification_matplot" and "correlation" plots. By default, each plot is printed to a new plotting area.
sig_correlation	The “significance-level” for plotting the correlation between variables. Note that this is an estimate of a posterior probability and not a significance level as defined in frequentist statistics. Default value: NULL (all correlations are plotted).
confidence	Confidence level(s) for plotting the Highest Density Interval(s) (as shown via what = 2). Default: confidence = 0.95.
...	ignored.

Details

When the BIC values are plotted, a number indicates the most probable number of “alive” clusters. The pairwise scatterplots (what = "classification_pairs") are created using the coordProj function of the mclust package. The what = "correlation" is plotted using the corrplot package. Note that the what = "classification_matplot" plots the original data (before scaling and centering). On the other hand, the option what = "classification_pairs" plots the centered and scaled data.

Author(s)

Panagiotis Papastamoulis

References

Luca Scrucca and Michael Fop and Thomas Brendan Murphy and Adrian E. Raftery (2017). mclust 5: clustering, classification and density estimation using Gaussian finite mixture models. The R Journal, 8(1): 205–233.

Taiyun Wei and Viliam Simko (2017). R package "corrplot": Visualization of a Correlation Matrix (Version 0.84). Available from https://github.com/taiyun/corrplot

fabMix documentation built on May 29, 2024, 2:53 a.m.