Skip to contents

Superclasses of GeoSOM

Source: R/geosom_helper.R
geosom_superclass.Rd

It is common to further cluster the GeoSOM map into superclasses, groups of cells with similar profiles. This is done using classic clustering algorithms on the map’s prototypes.Remeber to use it when the non-spatial variable is more than 1.

Usage

geosom_superclass(
  gsom,
  k,
  method = "pam",
  hmethod = "complete",
  bindcoord = FALSE
)

Arguments

gsom

A kohonen object,get from geosom().

k

The number of superclasses.

method

(optional)The clustering algorithms used on the map’s prototypes,two methods are implemented in spEcula, PAM(k-medians) and hierarchical clustering.When method is pam,PAM (k-medians) is used,otherwise hierarchical clustering.Default is pam.

hmethod

For hierarchicical clustering, the clustering method, by default "complete". See the stats::hclust documentation for more details.

bindcoord

Does the cluster of superclass in GeoSOM consider spatial coordination. Defaul is FALSE.

Value

A numeric vector representing the superclass.

Author

Wenbo Lv lyu.geosocial@gmail.com

Examples

data(pmc)
set.seed(2004)
g = geosom(data = pmc, coords = c("centroidx","centroidy"),
wt = 3,grid = geosomgrid(6,10),normalize = TRUE)
g_superclass = geosom_superclass(g,12)
g_superclass
#>  [1]  1  2  1  3  4  5  3  1  1  3  4  4  6  3  3  4  7  7  6  6  6  4  7  7  6
#> [26]  8  8  9  9  7  6  6  8  9  9  9  1  8  8 10 10  9 11 12 10 10 10  9 12 12
#> [51] 10 10  9  9  8 10 10 10  9  9