— Data stream mining research has gained importance in the recent years due to the generation of vast amount of data streams by many applications. Transactional data streams are characterized by high dimensionality and high cardinality. The transactional data streams arrive at a very high speed in an unbounded form. Clustering is an important core data mining activity that provides valuable insights into the data being processed. Clustering Transactional stream is a highly challenging activity as it is bound to single pass constraint as well as memory and CPU constraints. In this context an efficient algorithm is proposed to cluster the transactional data streams. The proposed algorithm accounts for the resource constraints. Extensive experimental analysis of the proposed algorithm on the real and synthetic data demonstrates the scalability and efficiency of the algorithm.
— Cluster histogram, data streams, data stream clustering, resource adaptation, sliding window model.
J. Chandrika is with Dept. of Computer Science and Engineering, MCE, Hassan, India (e-mail; email@example.com).
K. R. Ananda Kumar is with Department of Computer Science, SJBIT, Bangalore (e-mail:firstname.lastname@example.org).
Cite: J. Chandrika and K. R. Ananda Kumar, " A Novel Approach for Clustering Categorical Data Streams," International Journal of Innovation, Management and Technology vol. 4, no. 5, pp. 486-489, 2013.