Data Clustering Algorithms: Experimentation and Comparison

Due to increasing databases of all kinds, clustering has become one of the most essential tasks to classify the data. Clustering means to group or divide the data points based on their similarity to each other. Clustering can be stated as an unsupervised data mining technique that describes the nature of datasets. The main objective of data clustering is to obtain groups of similar entities. There are various methods of clustering such as hierarchical, partition-based, method-based, grid-based, and model-based. This paper provides a detailed study about clustering, its working processes. Along with basic information, detailed information about validity measures required to evaluate algorithms is discussed. This paper reviews clustering algorithms like K-Means, Agglomerative, and DBSCAN. A tabular comparison of algorithms is represented to acquire in-depth knowledge. The results obtained after experimenting with algorithms are also been discussed at the end.