Applicative implementation of D-stream clustering algorithm for the real-time data of telecom sector

In the recent past, telecommunication industry has gone through tremendous growth. It has resulted in huge production of data on daily basis for the telecom companies. Now it is a challenge for the telecom operators to manage huge amount of data and then use this data for decision and policy making processes. The data generated in telecom industry is in the form of data stream as it is continuously being generated. So we need such a clustering algorithm which can perform well with streams or continuous data. At the same time we also need to detect outliers and erroneous data. Clusters are not necessarily of globular shape as we don't have prior knowledge of number of clusters and their shape. To address all the above mentioned problems we have used D-Stream clustering algorithm in our implementation to get desired results. This paper, discusses the algorithm and implementation of DStream algorithm along with its experimental results on synthetically generated telecommunication data.