Hands-On Machine Learning with C++

Distance-based clustering algorithms

The most known representatives of this family of methods are the k-means and k-medoids algorithms. They take the k input parameter and divide the data space into k clusters, such that the similarity between objects in one cluster is maximal. Also, they minimize the similarity between objects of different clusters. The similarity value is calculated as the distance from the object to the cluster center. The main difference between these methods lies in the way the cluster center is defined.

With the k-means algorithm, the similarity is proportional to the distance to the cluster center of mass. The cluster center of mass is the average value of cluster objects' coordinates in the data space. The k-means ...