Practical Robust Statistical Methods via R

M Mächler

Abstract:The R language and environment for statistical computing and graphics has become the de facto standard for most of professional statisticians. Some robust methodology has been part of S, the precursor of R, since the 1980s. Further, in 2006, a considerable group of research experts in robust statistics had decided to join forces creating a dedicated R package providing the most basic and important robust methodology, notably covering Maronna et al.(2006), and the R package robustbase ((?)) was created. Not much later, when CRAN task views (CTV) were initiated as a human expert guides to the exponentially growing CRAN R packages, the “Robust” CTV ((?)) was created and listed dozens of R packages already in 2007 and has been listing 50 packages since the end of 2014.Our presentation will emphasize on use-cases of regression methods—in our experience used for roughly 90% of all data analysis …