[R-pkgs] New package pivmet: relabelling and K-means seeding via pivotal methods

legidi m@ili@g off u@its@it legidi m@ili@g off u@its@it
Fri Nov 23 13:13:27 CET 2018


Dear R users,

I am glad to announce the release (version 0.1) of the pivmet package,  
which proposes some pivotal methods in order to:


* undo the label switching problem which naturally arises during the  
MCMC sampling in Bayesian mixture models [pivotal relabelling] (Egidi  
et al. 2018a)

* initialize the K-means algorithm aimed at obtaining a good  
clustering solution [pivotal seeding] (Egidi et al. 2018b)

The package includes two vignettes for easing its use and is here available:

https://cran.r-project.org/web/packages/pivmet/index.html

and developed in github at:

https://github.com/LeoEgidi/pivmet


Here are the two referred articles:

Egidi et al. (2018a)
https://link.springer.com/article/10.1007/s11222-017-9774-2

Egidi et al. (2018b)
https://www.researchgate.net/profile/Leonardo_Egidi/publication/326225330_K-means_seeding_via_MUS_algorithm_-_Inizializzazione_del_K-means_tramite_l%27algoritmo_MUS/links/5b3f2c2caca27207851c7865/K-means-seeding-via-MUS-algorithm-Inizializzazione-del-K-means-tramite-lalgoritmo-MUS.pdf


All the best

Leonardo Egidi
Postdoctoral researcher, University of Trieste



More information about the R-packages mailing list