Techniques for Data Stream Clustering: Comparative Analysis

Abstract

 In the fast growing world applications are generating data in enormous volumes called data streams. Data stream is imaginably large, continual, rapid flow of information and in data mining the important tool is called clustering, hence data stream clustering (DSC) can be said as active research area. Recent attention of data stream clustering is through the applications that contain large amounts of streaming data. Data stream clustering is used in many areas such as weather forecasting, financial transactions, website analysis, sensor network monitoring, e-business, telephone records and telecommunications. In case of data stream clustering most popularly used heuristic is K-means and other algorithms like K-medoids and the popular BIRCH are developed. The aim of the abstract is to review the developments and trends of data stream clustering methods and analyze typical DSC algorithms proposed in recent years, such as BIRCH, STREAM, DSTREAM and some more algorithms.