Title: An effective hashtable-based approach for incrementally mining closed frequent itemsets using sliding windows

Authors: M. Jeya Sutha; F. Ramesh Dhanaseelan

Addresses: Department of Computer Applications, St. Xavier's Catholic College of Engineering, Anna University, Chunkankadai – 629003, K.K. Dist., Tamil Nadu, India ' Department of Computer Applications, St. Xavier's Catholic College of Engineering, Anna University, Chunkankadai – 629003, K.K. Dist., Tamil Nadu, India

Abstract: Online mining of closed frequent itemsets over streaming data is one of the important problems in mining data streams. In this paper, we propose a new algorithm called 'CFI-StreamSW' (mining closed frequent itemsets over data streams using sliding window), for mining the set of closed frequent itemsets. An effective hash table based approach is followed where two tables are used; one for storing all the items in the transactions and another for closed frequent itemsets. Thus, it does not store any other intermediate nodes or even frequent nodes. Experiments show that the proposed algorithm runs faster and consume less memory than existing algorithms 'NewMoment' and 'MWFP-SW' for mining closed frequent itemsets over recent data streams.

Keywords: data mining; closed frequent itemsets; sliding windows; SWs; landmark windows; damped windows; data streams; incremental mining; hash table; stream mining; single pass mining.

DOI: 10.1504/IJDMMM.2016.081252

International Journal of Data Mining, Modelling and Management, 2016 Vol.8 No.4, pp.382 - 404

Received: 27 Dec 2014
Accepted: 09 Aug 2015

Published online: 29 Dec 2016 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article