A Survey on Optimization Approaches to K-Means Clustering using Simulated Annealing

Abstract: Clustering is one of the fastest growing research areas because of availability of huge amount of data. It models data into the clusters. Data modelling puts clustering in a historical perspective rooted in statistics, mathematics, and numerical analysis. From a machine learning perception clusters correspond to hidden patterns, the exploration for clusters is unsupervised learning, the resultant system represents a data model. There are many techniques for clustering of data based on similarity. K-Means is one of the simplest unsupervised learning methods among all partitioning based clustering methods. It classifies a set of data objects in clusters. All the data objects are placed in a cluster having centroid nearest to that data object. After processing the data objects centroids are recalculated, and the whole process is repeated. This paper presents a brief estimation of the existing body of work that employs simulated annealing approach to improve upon the k-means clustering process