OPTICS Algorithm

Ordering points to identify the clustering structure, or OPTICS, is an algorithm for finding density-based clusters in spatial data, presented in 1999 by Mihael Ankerst, Markus Breunig, Hans-Peter Kriegel and Jorg Sander. Its basic idea is similar to DBSCAN, but it addresses one of that algorithm's major weaknesses, the difficulty of detecting meaningful clusters in data of varying density, by linearly ordering the points of the database so that spatially closest points become neighbors in the ordering and by storing, for each point, a special distance representing the density that must be accepted for a cluster so that both points belong to it, a structure that can be represented as a dendrogram.