Clustering

Clustering is an unsupervised learning task that organizes data into groups based on a chosen similarity measure. Algorithms such as k-means, hierarchical clustering, DBSCAN, and Gaussian mixture models make different assumptions about cluster shape, density, number, and overlap. Clustering can help explore customer behavior, documents, images, biological data, or anomalies when predefined labels do not exist. The output is not an objective discovery of natural categories: feature choices, scaling, distance metrics, initialization, and algorithm settings can substantially change the groups. Useful clusters should be stable, interpretable, and validated against the practical question rather than accepted solely because a visualization looks separated.