Consolidated study & analysis of different clustering techniques for data streams

Recently a data mining particularly a stream data has been drawing widespread attentions of the researchers and the experts around the globe and a huge number of interesting research outcomes have been achieved. The scope of data generation and analytics has increased exponentially due to increase in the data transfer rate, integration and connection of the digitalized systems. Today most of the device generates data which are huge in volume, non-stationary and fast changing. As the data available is huge in size there is requirement of clustering in order to enhance the quality of data that will provide desired outcome. Clustering is the procedure of making a group of abstract objects into classes of similar objects. Clustering will increase data processing speed which is most important in data mining. So clustering has become challenging field of research because of their special requirements of its applications. In data stream mining the clustering possess very important role. Stream data are the types of data sets which are massive, complex, fast changing and potentially infinite. We cannot gather each element of a stream since stream flows so quick and changes so rapidly. Sometimes there is requirement of many applications to group the data based on its characteristics so clustering is done on such types of data. In this paper we have reviewed typical requirements of clustering in data mining, major challenges and issues of the clustering methods, different techniques of clustering for data mining with basic algorithm of each of these methods with their advantages and disadvantages and methods used for stream data.