Model-Based Sliced Inverse Regression

An R package for dimension reduction based on finite Gaussian mixture modeling of inverse regression.


News

msir 1.3.2 (2019-03)

  • Update vignette.
  • Added website for inclusion in mclust-org.
  • Fix a bug in msir.slices().
  • Fix a bug on recycling array of length 1.
  • rgl package is no longer a dependency but only suggested.

msir 1.3.1 (2016-04)

  • Fix issues related to NAMESPACE.

msir 1.3 (2013-02)

  • Removed MASS as suggested package since now it is not needed for examples in help pages.
  • Fixed comments and notes from package check.

msir 1.2 (2012-08)

  • Compatibility bug fixes with new version 4.0 of mclust package.
  • Add verbose argument to summary.msir().

msir 1.1 (2012-05)

  • New mvdnorm() function which should be more stable in the case of near singularity of covariance matrix.
  • Added byte-code compiling in the DESCRIPTION file.
  • Corrected msir.components(), now it preserves the order of slices.
  • Introduced the possibility to control EM algorithm.
  • Bug correction and change in models' name in msir.regularizedSigma().
  • The 'msir' object returns the normalized basis vectors in component $basis and raw eigenvectors in component $evectors.
  • Update reference to CSDA paper.
  • Bugs fix.

msir 1.0 (2012-05)

  • Initial release submitted to CRAN.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("msir")

1.3.2 by Luca Scrucca, 5 months ago


https://mclust-org.github.io/msir


Browse source code at https://github.com/cran/msir


Authors: Luca Scrucca [aut, cre]


Documentation:   PDF Manual  


GPL (>= 2) license


Imports mclust, stats, utils, graphics, grDevices

Suggests knitr, rmarkdown, rgl


See at CRAN