Indian Institute of Technology Roorkee
Clustering of data streams is an important issue in data mining. A large number of algorithms exist for clustering data streams but most of these algorithms give equal weights to all the dimensions of the data stream. Some of the dimensions of the data stream may play important role in clustering while some may be just useless. In this paper, the authors introduce a density based algorithm in which the dimensions of the streaming data are assigned weights according to the importance of that dimension in the process of clustering. This approach is highly effective and efficient.