pkg/man/clustering.Rd

   1 % Generated by roxygen2: do not edit by hand
   2 % Please edit documentation in R/clustering.R
   3 \name{clustering}
   4 \alias{clustering}
   5 \alias{clusteringTask1}
   6 \alias{computeClusters1}
   7 \alias{computeClusters2}
   8 \alias{clusteringTask1}
   9 \alias{clusteringTask2}
  10 \alias{computeClusters1}
  11 \alias{computeClusters2}
  12 \title{Two-stage clustering, withing one task (see \code{claws()})}
  13 \usage{
  14 clusteringTask1(indices, getContribs, K1, nb_series_per_chunk,
  15   ncores_clust = 1, verbose = FALSE, parll = TRUE)
  16
  17 clusteringTask2(medoids, K2, getRefSeries, nb_ref_curves, nb_series_per_chunk,
  18   ncores_clust = 1, verbose = FALSE, parll = TRUE)
  19
  20 computeClusters1(contribs, K1, verbose = FALSE)
  21
  22 computeClusters2(distances, K2, verbose = FALSE)
  23 }
  24 \arguments{
  25 \item{indices}{Range of series indices to cluster in parallel (initial data)}
  26
  27 \item{getContribs}{Function to retrieve contributions from initial series indices:
  28 \code{getContribs(indices)} outpus a contributions matrix}
  29
  30 \item{K1}{Number of super-consumers to be found after stage 1 (K1 << N)}
  31
  32 \item{nb_series_per_chunk}{(~Maximum) number of series in each group, inside a task}
  33
  34 \item{ncores_clust}{"OpenMP" number of parallel clusterings in one task}
  35
  36 \item{verbose}{Level of verbosity (0/FALSE for nothing or 1/TRUE for all; devel stage)}
  37
  38 \item{parll}{TRUE to fully parallelize; otherwise run sequentially (debug, comparison)}
  39
  40 \item{medoids}{big.matrix of medoids (curves of same length as initial series)}
  41
  42 \item{K2}{Number of clusters to be found after stage 2 (K2 << K1)}
  43
  44 \item{getRefSeries}{Function to retrieve initial series (e.g. in stage 2 after series
  45 have been replaced by stage-1 medoids)}
  46
  47 \item{nb_ref_curves}{How many reference series? (This number is known at this stage)}
  48
  49 \item{contribs}{matrix of contributions (e.g. output of \code{curvesToContribs()})}
  50
  51 \item{distances}{matrix of K1 x K1 (WER) distances between synchrones}
  52 }
  53 \value{
  54 For \code{clusteringTask1()} and \code{computeClusters1()}, the indices of the
  55   computed (K1) medoids. Indices are irrelevant for stage 2 clustering, thus
  56   \code{computeClusters2()} outputs a big.matrix of medoids
  57   (of size limited by nb_series_per_chunk)
  58 }
  59 \description{
  60 \code{clusteringTask1()} runs one full stage-1 task, which consists in
  61   iterated stage 1 clustering (on nb_curves / ntasks energy contributions, computed
  62   through discrete wavelets coefficients).
  63   \code{clusteringTask2()} runs a full stage-2 task, which consists in synchrones
  64   and then WER distances computations, before applying the clustering algorithm.
  65   \code{computeClusters1()} and \code{computeClusters2()} correspond to the atomic
  66   clustering procedures respectively for stage 1 and 2. The former applies the
  67   clustering algorithm (PAM) on a contributions matrix, while the latter clusters
  68   a chunk of series inside one task (~max nb_series_per_chunk)
  69 }