LeaDen-Stream: A Leader Density-Based Clustering Algorithm over Evolving Data StreamReport as inadecuate




LeaDen-Stream: A Leader Density-Based Clustering Algorithm over Evolving Data Stream - Download this document for free, or read online. Document in PDF available to download.

Clustering evolving data streams isimportant to be performed in a limited time with a reasonable quality. Theexisting micro clustering based methods do not consider the distribution ofdata points inside the micro cluster. We propose LeaDen-Stream Leader Density-basedclustering algorithm over evolving data Stream, a density-basedclustering algorithm using leader clustering. The algorithm is based on atwo-phase clustering. The online phase selects the proper mini-micro ormicro-cluster leaders based on the distribution of data points in the microclusters. Then, the leader centers are sent to the offline phase to form finalclusters. In LeaDen-Stream, by carefully choosing between two kinds of microleaders, we decrease time complexity of the clustering while maintaining thecluster quality. A pruning strategy is also used to filter out real data fromnoise by introducing dense and sparse mini-micro and micro-cluster leaders. Ourperformance study over a number of real and synthetic data sets demonstratesthe effectiveness and efficiency of our method.

KEYWORDS

Evolving Data Streams; Density-Based Clustering; Micro Cluster; Mini-Micro Cluster

Cite this paper

Amini, A. and Wah, T. 2013 LeaDen-Stream: A Leader Density-Based Clustering Algorithm over Evolving Data Stream. Journal of Computer and Communications, 1, 26-31. doi: 10.4236-jcc.2013.15005.





Author: Amineh Amini, Teh Ying Wah

Source: http://www.scirp.org/



DOWNLOAD PDF




Related documents