An efficient algorithm for mining toprank-k frequent patterns

Thu-Lan Dam, Kenli Li, Philippe, Fournier-Viger & Quang-Huy Duong

An efficient algorithm for mining toprank-k frequent patterns

Thu-Lan Dam, Kenli Li, Philippe, Fournier-Viger & Quang-Huy Duong

The International Journal of Artificial, Intelligence, Neural Networks, and Complex Problem-Solving Technologies (ISSN 0924-669X, Volume 45, Number 1)

2016

Abstract

Mining top-rank-k frequent patterns is a popular data mining task, which consists of discovering the patterns in a transaction database that belong to the k first ranks in terms of support. Although, several algorithms have been proposed for this task, it remains computationally expensive. To address this issue, this paper proposes a novel algorithm named BTK. It relies on a novel tree structure named TB-tree to store crucial information about frequent patterns. Moreover, BTK employs a new B-list structure to store information about patterns, and relies on subsume indexes to reduce the search space and speed up the discovery of top-rank-k frequent patterns. BTK also uses an early pruning strategy and an effective threshold raising mechanism. Additionally, BTK introduces two efficient procedures for respectively generating subsume indexes and intersecting B-lists. Extensive experiments were conducted on several datasets to evaluate the efficiency of the proposed algorithm. Results show that BTK is highly efficient and competitive.

Citation

Thu-Lan Dam, Kenli Li, Philippe, Fournier-Viger & Quang-Huy Duong, “An efficient algorithm for mining toprank-k frequent patterns”, The International Journal of Artificial, Intelligence, Neural Networks, and Complex Problem-Solving Technologies (ISSN 0924-669X, Volume 45, Number 1), 2016.

Full paper Pdf: https://doi.org/10.1007/s10489-015-0748-9

  • Thứ Sáu, 10:57 27/04/2018

Tags: