permute

Gavin Simpson

What is permute?

permute is an R package that provides functions to generate restricted and unrestricted permutations. The package was originally conceived as an extension for the vegan package, essentially as internal functions for that package. Since then permute has undergone a number of revisions and is now a separate package in its own right, allowing permute’s functions to be used within other R packages.

The restricted permutations in permute are modelled on the permutations schemes of Cajo ter Braak available in the Canoco software since version 3.1.

The main functions in permute are shuffle() and shuffleSet() which return a single permutation or a set of n permutations, and how(), which is used to set up the permutation design.

The package comes with a large number of extractor functions to access aspects of the permutation design. It is recommended that these be used rather than subsetting the object return by how() as the internal representation of the permutation design may change.

Features

Large range of permutation schemes
- Unrestricted (i.e. randomisation)
- Line transects or time series
- Spatial grids
- Split-plot designs; whole-plot and split plots can be any of the above permutation types or held fixed
- Blocking factors (samples never permuted between blocks, only within)
Allows complete enumeration of all possible permutations (within machine and memory restrictions)
Checking of permutation designs (balance within plots where required, etc)
Calculation of maximum number of permutations for current design