Skip to contents

Model fitting

EM-type estimators for univariate Gaussian mixture models. All fitters share the same interface and return converged/iter alongside the parameter estimates.

robustGMM()
Fitting robust Gaussian Mixture Model
vanillaGMM()
Standard Gaussian Mixture Model via EM
uniformNoiseGMM()
Gaussian Mixture Model with a Uniform Noise Component

Tuning and objective functions

The beta-divergence objective and the leave-one-out Cramer-von Mises criterion for choosing the robustness tuning parameter.

loo_cvm_div()
Leave-One-Out Cramer-von Mises Divergence
beta_likelihood()
Beta-Likelihood of Gaussian Mixture Model
beta_responsibility()
Generalized (Beta) Responsibility
b_beta()
B-beta function of GMM model

Normal mixture utilities

dnormix()
Density of normal mixture
pnormix()
Cumulative distribution function of normal mixture
rnormix()
Generate random samples from normal mixture
lse()
Log-Sum-Exp