Comparison of Clustering Algorithms: K-Means, DBSCAN and Ward’s method

Several clustering algorithms have been introduced to literature in the last 10 years. Clustering methods usage depends on their complexity, the amount of data, the purpose of clustering and the predefined parameters. This case study, presents three of the most used clustering algorithms, K-means, DBSCAN and Ward’s method.

K-means

K-means belongs to partitioning spatial clustering algorithms. It is a frequently used clustering method and it is one of the simplest unsupervised learning algorithms. K-means defines clusters by partitioning all observations into groups, in which each observation belongs to the group with the nearest mean. The algorithm operates in iterations until the sum of squares from points to the assigned cluster centres is minimised. The end result of k-means algorithm is the partitioning of the data space into Voronoi cells. Continue reading “Comparison of Clustering Algorithms: K-Means, DBSCAN and Ward’s method”