plotGMM Tools for Visualizing Gaussian Mixture Models

ATTENTION: The package (including it's name) is under active development with significant updates coming shortly. Thus, the applications and examples below still apply for the current stable release (v 0.2.1), but will soon be significantly updated and expanded. Stand by for developments and news!

In collaboration with Fong Chan (Achilles Therapeutics) and Lu Zhang (Emory University), the latest release (v0.2.1) of plotGMM includes substantial updates with easy-to-use tools for visualizing output from univariate Gaussian mixture models:

1. plot_GMM: The main function of the package, plot_GMM allows the user to simply input the name of a mixEM class object (from fitting a Gaussian mixture model (GMM) using the mixtools package), as well as the number of components, k, that were used in the GMM fit. The result is a clean ggplot2 object showing the density of the data with overlaid mixture weight component curves.

2. plot_cut_point: Gaussian mixture models (GMMs) are not only used for uncovering clusters in data, but are also often used to derive cut points, or lines of separation between clusters in feature space (see the Benaglia et al. 2009 reference in the package documentation for more). The plot_cut_point function plots data densities with the overlaid cut point (the mean of the calculated mu) from mixEM class objects, which are GMM's fit using the mixtools package.

3. plot_mix_comps: This is a custom function for users interested in manually overlaying the components from a Gaussian mixture model. This allows for clean, precise plotting constraints, including mean (mu), variance (sigma), and mixture weight (lambda) of the components. The function superimposes the shape of the components over a ggplot2 class object. Importantly, while the plot_mix_comps function is used in the main plot_GMM function in our plotGMM package, users can use the plot_mix_comps function to build their own custom plots.

Installation

Dev version: devtools::install_github("pdwaggoner/plotGMM")

CRAN (v0.2.1) release: install.packages("plotGMM"); library(plotGMM)

Plotting GMMs using plot_GMM

# Fit a GMM using EM
set.seed(576)
mixmdl <- mixtools::normalmixEM(iris$Petal.Length, k = 2)

# Plot the density with overlaid mixture weight curves
plot_GMM(mixmdl, 2)

Plotting cut points from GMMs using plot_cut_point

# Fit a GMM using EM via mixtools
set.seed(576)
mixmdl <- mixtools::normalmixEM(faithful$waiting, k = 2)

# Option 1.1: Plot with amerika palette
plot_cut_point(mixmdl, plot = TRUE,
                color = "amerika")

# Option 1.2: Plot with wesanderson palette
plot_cut_point(mixmdl, plot = TRUE,
                color = "wesanderson")

# Option 1.3: Plot with grayscale color, overriding default labels
plot_cut_point(mixmdl, plot = TRUE,
                color = "grayscale") +
  ggplot2::labs(x = "Waiting Time (Minutes)",
                title = "Cutpoint from GMM for Old Faithful Waiting Time")

# Option 2: Cutpoint value only
plot_cut_point(mixmdl, plot = FALSE) # 67.35299

Use plot_mix_comps for a custom plot manually overlaying component curves

library(plotGMM)
library(magrittr)
library(ggplot2)
library(mixtools)

# Fit a GMM using EM
set.seed(576)
mixmdl <- normalmixEM(faithful$waiting, k = 2)

# Plot mixture components using the `plot_mix_comps` function
data.frame(x = mixmdl$x) %>%
ggplot() +
geom_histogram(aes(x, ..density..), binwidth = 1, colour = "black",
                 fill = "white") +
   stat_function(geom = "line", fun = plot_mix_comps,
                 args = list(mixmdl$mu[1], mixmdl$sigma[1], lam = mixmdl$lambda[1]),
                 colour = "red", lwd = 1.5) +
   stat_function(geom = "line", fun = plot_mix_comps,
                 args = list(mixmdl$mu[2], mixmdl$sigma[2], lam = mixmdl$lambda[2]),
                 colour = "blue", lwd = 1.5)