K-medoids

K-medoids is a clustering algorithm that selects actual data points as cluster centers.

It is widely used in clustering tasks where robustness to outliers is important. K-medoids operates similarly to K-means but selects actual data points as cluster centers (medoids) rather than using the mean of the points in a cluster.

The algorithm follows these steps:

1. On each round, the input data item that is closest to the geometric center of each cluster is selected to be used as the new cluster center for the subsequent round.
2. The two subtypes, Partitioning around medoids and Voronoi iteration, are simply different ways of achieving this mathematically.

Since K-medoids selects actual data points as cluster centers, it is less sensitive to the presence of outliers, which can significantly affect the mean used in K-means. Hence, it provides more stable and reliable clusters.