Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

compare_clusterings_configs

Create clustering configurations.


Description

Create configurations for compare_clusterings()

Usage

compare_clusterings_configs(
  types = c("p", "h", "f"),
  k = 2L,
  controls = NULL,
  preprocs = pdc_configs("preproc", none = list()),
  distances = pdc_configs("distance", dtw_basic = list()),
  centroids = pdc_configs("centroid", default = list()),
  no.expand = character(0L)
)

Arguments

types

Clustering types. It must be any combination of (possibly abbreviated): partitional, hierarchical, fuzzy, tadpole.

k

A numeric vector with one or more elements specifying the number of clusters to test.

controls

A named list of tsclust-controls. NULL means defaults. See details.

preprocs

Preprocessing configurations. See details.

distances

Distance configurations. See details.

centroids

Centroid configurations. See details.

no.expand

A character vector indicating parameters that should not be expanded between pdc_configs() configurations. See examples.

Details

Preprocessing, distance and centroid configurations are specified with the helper function pdc_configs(), refer to the examples in compare_clusterings() to see how this is used.

The controls list may be specified with the usual tsclust-controls functions. The names of the list must correspond to "partitional", "hierarchical", "fuzzy" or "tadpole" clustering. Again, please refer to the examples in compare_clusterings().

Value

A list for each clustering type, each of which includes a data frame with the computed and merged configurations. Each data frame has an extra attribute num.configs specifying the number of configurations.

Examples

# compare this with leaving no.expand empty
compare_clusterings_configs(
    distances = pdc_configs("d", dtw_basic = list(window.size = 1L:2L, norm = c("L1", "L2"))),
    centroids = pdc_configs("c", dba = list(window.size = 1L:2L, norm = c("L1", "L2"))),
    no.expand = c("window.size", "norm")
)

dtwclust

Time Series Clustering Along with Optimizations for the Dynamic Time Warping Distance

v5.5.10
GPL-3
Authors
Alexis Sarda-Espinosa
Initial release
2022-04-15

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.