Permutation Distribution Clustering

Permutation Distribution Clustering is a clustering method for time series. Dissimilarity of time series is formalized as the divergence between their permutation distributions. The permutation distribution was proposed as measure of the complexity of a time series.


V1.0.3; 28.09.2015

  • updated citation file and citation in docs
  • updated package's global imports to new CRAN policy

V1.0.2; 11.11.2014

  • bugfixes for resetting par() after plotting (plot.pdclust and rasterPlot)
  • removed from rasterPlot
  • fixed dataset "star.shapes.raw", which erroneously contained wrong images
  • updated citation information in docs

V1.0.1; 10.11.2014

  • calls to pdcDist with default arguments do not produce a warning anymore

V1.0; 16.10.2014

  • changed all function calls to camelcase naming convention, e.g., pdcDist() instead of pdc.dist().
  • pdclust objects inherit hclust class
  • fixed some problems with par() settings in plot()
  • added a tests directory
  • added rasterPlot for plotting results of shape clustering
  • added raw image data for complex.shape and star.shape data
  • changed print() of entropy heuristic object to sumamry() and added a new print() method
  • added proper NAMESPACE for S3 methods
  • added invisible() returns to prints and plots
  • various minor bugfixes

V0.5; 13.02.2014

  • added mds.plot() for multidimensional scaling including convex hulls of clusters
  • added string representation of pdclust dendrogram using str()
  • print() shows time-delay
  • added capability to plot p values for hypothesis tests on branch splits
  • added compatibility of loo1nn() with dissimilarity matrices

V0.4; 04.02.2014

  • added hellinger distance, squared hellinger distance, and symmetric alpha divergence
  • added shape signature tracer
  • added example datasets for shape signatures: complex.shapes and star.shapes
  • moved dependence of package lattice to suggestions

V0.3; 15.05.2012

  • drastic speed-up with fast C-implementation of codebook()
  • added support for multi-channel time-series

V0.2; 19.04.2012

  • implemented time-delayed embedding
  • drastic speed-up of codeword()-function
  • added function loo1nn()

V0.1; 19.03.2012

  • initial submission 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.


1.0.3 by Andreas M. Brandmaier, 6 years ago

Browse source code at

Authors: Andreas M. Brandmaier

Documentation:   PDF Manual  

Task views: Time Series Analysis

GPL (>= 3) license

Imports stats, utils, grDevices, graphics

Suggests plotrix, lattice

Imported by TSdist.

Depended on by TSclust.

See at CRAN