Matías Salibián-Barrera [Software]



  • A robust and sparse K-means clustering algorithm. An R package (called RSKC) to compute robust and sparse clusters based on K-means can be found in CRAN (a direct link is here). This work is based on Kondo, Y., Salibian-Barrera, M. and Zamar, R. H. (2012) A preprint is available at arXiv:1201.6082v1. Yumi's MSc thesis is available here.

  • An outlier robust fit for Generalized Additive Models - R package to compute a robust fit for Generalized Additive Models can be found in here. Joint work with Azadeh Alimadad.

  • Penalised S-regression splines - R code to compute penalised S-regression splines can be found here. Joint work with Tharmaratnam, K., Claeskens, G. and Croux, C. at K. U. Leuven.

  • Fast tau - MATLAB / OCTAVE and R code to compute the tau-regression estimator for linear regression can be found here. This is joint work with Gert Willems.

  • Globally robust confidence intervals for simple linear regression models - R code to compute globally robust confidence intervals for the slope of a simple linear regression model can be found here. This is joint work with Jorge Adrover.

  • Fast S - Plain R & S-PLUS code to compute the Fast-S estimator for linear regression can be found here. This algorithm is also implemented in the S- and MM-estimators in the robustbase package for R (see below). This is joint work with Prof. Victor Yohai.

  • roblm - The roblm R package for MM-regression estimators, version 0.6. The latest version of this package can be found in CRAN. This package is not longer maintained. All its functionality has been incorporated into the robustbase package (see below).

  • robustbase - An R package implementing state-of-the-art robust methods, particularly those described in the book Robust statistics, theory and methods, by Maronna, Martin and Yohai, Wiley, 2006. Available in CRAN. Currently maintained by Martin Maechler.

  • R package for the Linear Grouping Algorithm - [Van Aelst, S., Wang, X., Zamar, R., and Zhu, R. Linear grouping using orthogonal regression, Computational Statistics and Data Analysis (2006)]. It is able to use multiple CPUs if available and it implements the GAP statistic [Tibshirani, R., Walther, G. and Hastie, T. Estimating the number of clusters in a data set via the gap statistic, JRSS B, (2001)]. It is available here. This is joint work with Justin Harrington.






matias + stat - ubc - ca
Phone: (604) 822-3410
Fax: (604) 822-6960