RoCoCo - An R Package Implementing a Robust Rank Correlation Coefficient and a Corresponding Test

Classical rank correlation measures are designed for ordinal data, but they are not ideally suited for measuring rank correlation for numerical data that are perturbed by noise. We introduced a family of robust rank correlation measures on the basis of Goodman's and Kruskal's gamma, where the classical ordering of real numbers is replaced by some fuzzy ordering with smooth transitions - thereby ensuring that the correlation measure is continuous with respect to the data. The present R package implements this rank correlation measure along with a corresponding statistical test. Moreover, the package also implements the Gaussian rank correlation estimator and a corresponding statistical test.


The package is available through CRAN - The Comprehensive R Archive Network (click here to view the archive entry of the package). Therefore, the simplest way to install the package is to enter
into your R session. If, for what reason ever, you prefer to install the package manually, follow the instructions in the user manual.

The current version of the package is 1.1.1 (released 2014-02-27).


  1. User Manual: PDF (350KB; last updated on 2014-02-27)
  2. Reference Manual: PDF (103KB; last updated on 2014-02-27; generated automatically from package help pages)

Getting started

  1. To load the package, enter "library(rococo)" in your R session.
  2. To view the user manual, enter "vignette("rococo")".
  3. To do a first example, enter "example(rococo)".

Citing this package

If you use this package for research that is published later, you are kindly asked to cite it as follows:

U. Bodenhofer, M. Krone, and F. Klawonn (2012). Testing noisy numerical data for monotonic association. Inform. Sci., DOI: 10.1016/j.ins.2012.11.026.

U. Bodenhofer and F. Klawonn (2008). Robust rank correlation coefficients on the basis of fuzzy orderings: initial steps. Mathware Soft Comput. 15(1):5-20.

Change log

Version 1.1.1, released 2014-02-27:
  • adapted dependencies and linking to Rcpp version 0.11.0
  • cleared up package dependencies/li>
Version 1.1.0, released 2013-01-10:
  • estimation of p-values now based on relative frequencies
  • Gaussian rank correlation estimator and test added
  • corresponding adaptations of man pages and vignette
  • updated citation
  • moved vignette to "vignettes" according to the new guidelines
Version 1.0.1, released 2011-10-24:
  • improved estimation of p-values
  • added exact computation of p-values (only for 10 or less samples)
Version 1.0.0, released 2011-10-03:
first public release


For suggestions, bug reports, and other matters regarding the package, please contact