Clustering

The task of grouping a set of objects in such a way that objects in the same group are more similar to each other than to those in other groups.

How it works

Clustering algorithms find natural groupings in data without using labelled examples. K-Means iteratively assigns each point to the nearest cluster centroid and then moves the centroid to the mean of its assigned points until the assignments stabilise. Hierarchical clustering builds a tree of clusters by iteratively merging the two most similar clusters. Density-based approaches like DBSCAN find clusters as dense regions separated by sparse areas, naturally handling irregularly shaped groups and noise.

Why it matters

Clustering is essential for making sense of unlabelled data — one of the most common real-world situations. It powers customer segmentation, document organisation, anomaly detection, and the initial exploration of new datasets. In AI and ML pipelines, clustering embeddings from a language model is a common way to discover latent topics or group semantically similar content without manual labelling.