For EMCluster objects, change duplicate "X1" to "X2" in the plot title; typo in original release.

TODO for plotGMM:

• tests for cutpoints
• plotting options (e.g.. histograms, density plots, etc.)
• allow for non-mixEM GMM objects (i.e., interface with other packages for fitting GMMs
• simulations (?)

The function plot_cut_point calculates the optimal cut point by taking the mean of the calculated mu of the mixture model. There are two issues:

• This only makes sense for a 2-component mixture model. However, the function also allows models with k>=3 as argument.
• Even for a 2-component Gaussian mixture model, this is the optimal threshold only in case of a mixture model with equal priors and equal variances. In all other cases the equation is more complex and also has to take the priors and variances into account.

Hi

We preparing the next release of ggplot2 and our reverse dependency checks show that your package is failing with the new version. Looking into it we see that the example for plot_gmm() errors because the input to stat_function() returns NULL.

You can install the release candidate of ggplot2 using devtools::install_github('tidyverse/[email protected]') to test this out.

We plan to submit ggplot2 by the end of October and hope you can have a fix ready before then

Kind regards
Thomas

I was trying to visualize a two-component Gaussian mixture model with unequal covariance matrices for the two components. An error "Error in eigen(sigma) : non-square matrix in 'eigen'" is reported.
I think the reason is that mvnormalmixEM in mixtools gives two types of output for estimated mean and covariance matrix: vector/matrix when arbmean/arbvar is FALSE (i.e. mean and covariance in all components is equal), or list when the arbmean/arbvar is TRUE. So there are three cases: equal mean and unequal covariance, unequal mean and equal covariance, and unequal mean and covariance. The plot_mm could only handle the case with unequal mean and equal covariance.

TODO (e.g., BayesMix, mclust, etc.)

TODO