Skip to contents

This function take a data.frame and performs hierarchical clustering with contiguity constraints.

Usage

gtclust_knn(df, k = 3, method = "ward", scaling = "raw")

Arguments

df

an sf data.frame with polygons like features

k

number of nearest neighbors to take for building the graph (the graph will be symmetric so some points may have in fine more neighbors)

method

linkage criterion in ward (default) or average, median

scaling

default scaling of the features in zscore (default) or raw (i.e. no scaling)

Value

an hclust like object with additional slots

data

The numeric data (eventually scaled) used for the clustering

centers

The protoypes of each tree nodes

leafs_geometry

geometries of the dendrogram leafs as an sfc list

geotree

geometries of the dendrogram no-leafs node as an sfc list