Centroid Based Clustering Algorithms- A Clarion Study

Abstract — The main motto of data mining techniques is to generate user-centric reports basing on the business requirements. The advent increase in thirst for knowledge discovery has increased the need of robust algorithms in the process of knowledge discovery. Mining in general is termed as intrinsic methodology of discovering interesting, formerly unknown data patterns. Clustering can be termed as a set-grouping task where similar objects are being grouped together. Clustering, a primitive anthropological method is the vital method in exploratory data mining for statistical data analysis, machine learning, and image analysis and in many other predominant branches of supervised and unsupervised learning. Cluster algorithms can be categorized based on the cluster models available depending on type of data we try to analyse. This paper focuses on different centroid based algorithms on par numerical and categorical data