Clustering Algorithms

In this chapter, key aspects of clustering analysis are presented. Clustering is compared to regression. Types of clustering are distinguished. Details of centroid-based clustering, density-based clustering, distribution-based clustering, and hierarchical or connectivity-based clustering are provided. Centroid-based clustering groups data points into clusters based on their proximity to cluster centroids (mean or medoid). Density-based clustering identifies clusters as areas of high density separated by regions of low density. Distribution-based clustering assumes that data points belong to different probability distributions (e.g., Gaussian), and identifies clusters based on these distributions. Hierarchical clustering builds a hierarchy of clusters, represented as a dendrogram. Clusters are merged (agglomerative) or split (divisive) based on distance or similarity. A comparison is made between these clustering algorithms.