Deep Gaussian mixture models as proposed by Viroli and McLachlan (2019) <doi:10.1007/s11222-017-9793-z> provide a generalization of classical Gaussian mixtures to multiple layers. Each layer contains a set of latent variables that follow a mixture of Gaussian distributions. To avoid overparameterized solutions, dimension reduction is applied at each layer by way of factor models.
|Imports:||mvtnorm, corpcor, mclust|
|Author:||Cinzia Viroli, Geoffrey J. McLachlan|
|Maintainer:||Suren Rathnayake <surenr at gmail.com>|
|License:||GPL (≥ 3)|
|CRAN checks:||deepgmm results|
|Windows binaries:||r-devel: deepgmm_0.1.62.zip, r-release: deepgmm_0.1.62.zip, r-oldrel: deepgmm_0.1.62.zip|
|macOS binaries:||r-release (arm64): deepgmm_0.1.62.tgz, r-oldrel (arm64): deepgmm_0.1.62.tgz, r-release (x86_64): deepgmm_0.1.62.tgz, r-oldrel (x86_64): deepgmm_0.1.62.tgz|
|Old sources:||deepgmm archive|
Please use the canonical form https://CRAN.R-project.org/package=deepgmm to link to this page.