DistributionOptimization: Distribution Optimization

Fits Gaussian Mixtures by applying evolution. As fitness function a mixture of the chi square test for distributions and a novel measure for approximating the common area under curves between multiple Gaussians is used. The package presents an alternative to the commonly used Likelihood Maximization as is used in Expectation Maximization. The algorithm and applications of this package are published under: Lerch, F., Ultsch, A., Lotsch, J. (2020) <doi:10.1038/s41598-020-57432-w>. The evolution is based on the 'GA' package: Scrucca, L. (2013) <doi:10.18637/jss.v053.i04> while the Gaussian Mixture Logic stems from 'AdaptGauss': Ultsch, A, et al. (2015) <doi:10.3390/ijms161025897>.

Package details

AuthorFlorian Lerch, Jorn Lotsch, Alfred Ultsch
MaintainerFlorian Lerch <lerch@mathematik.uni-marburg.de>
LicenseMIT + file LICENSE
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:

Try the DistributionOptimization package in your browser

Any scripts or data that you put into this service are public.

DistributionOptimization documentation built on Feb. 12, 2020, 5:57 p.m.