geocmeans: An R package for spatial fuzzy c-means
Jérémy Gelb · The Journal of Open Source Software · 2023
Unsupervised classification methods like k-means or the Hierarchical Cluster Analysis (HCA) are widely used in geography even though they are not well suited for spatial data [Romary et al. (2015);] because they do not consider space.Yet, recent development has been proposed to include the geographical dimension into clustering.As an example, ClustGeo (Chavent et al., 2018) is a spatial extension of the HAC, available in the R package with the same name.We present here the R package geocmeans, proposing several spatial extensions of the Fuzzy C-Means (FCM) algorithm to complete this growing toolbox with a fuzzy approach.The package provides also several helper functions to assess and compare quality of classifications, select appropriate hyperparameters, and interpret the final groups.