Clustering Using OPTICS – Towards Data Science

Clustering is a powerful unsupervised knowledge discovery tool used today, which aims to segment your data points into groups of similar features. However, each algorithm is pretty sensitive to the parameters. Similarity based techniques (K-means, etc) are tasked with designating how many clusters exist, while hierarchical usually require manual intervention to decide when to assign finished clusters. The most common density based approach, DBSCAN, requires only two parameters pertaining to how it defines its "Core Points", but finding the parameters can often be an extremely difficult task. It also will not be able to find clusters of differing densities. There is a relative of DBSCAN, called OPTICS (Ordering Points to Identify Cluster Structure), that invokes a different process.