Abstract

Clustering has been widely used as a tool to group multivariate observations that have similar characteristics. However, most attempts at formulating a method to group similar multivariate observations while taking into account their spatial location are relatively ad hoc and do not account for the underlying spatial structure of the variables measured [12, 13, 14]. This paper proposes a method to spatially cluster similar observations based on the likelihood function. The geographic or spatial location of the observations can be incorporated into the likelihood of the multivariate normal distribution through the variance-covariance matrix. The variance-covariance matrix can be computed using any specific spatial covariance structure. Therefore, observations within a cluster which are spatially close to one another will have a larger likelihood than those observations which are not close to each other. This results in observations which are similar and spatially close to one another being placed into the same cluster.

Keywords

spatial clustering, geostatistics, multivariate likelihood, spherical covariance

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
Apr 27th, 1:30 PM

SPATIAL CLUSTERING USING THE LIKELIHOOD FUNCTION

Clustering has been widely used as a tool to group multivariate observations that have similar characteristics. However, most attempts at formulating a method to group similar multivariate observations while taking into account their spatial location are relatively ad hoc and do not account for the underlying spatial structure of the variables measured [12, 13, 14]. This paper proposes a method to spatially cluster similar observations based on the likelihood function. The geographic or spatial location of the observations can be incorporated into the likelihood of the multivariate normal distribution through the variance-covariance matrix. The variance-covariance matrix can be computed using any specific spatial covariance structure. Therefore, observations within a cluster which are spatially close to one another will have a larger likelihood than those observations which are not close to each other. This results in observations which are similar and spatially close to one another being placed into the same cluster.