Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10321827 | Expert Systems with Applications | 2015 | 26 Pages |
Abstract
Frequent closed patterns (FCPs), a condensed representation of frequent patterns, have been proposed for the mining of (minimal) non-redundant association rules to improve performance in terms of memory usage and mining time. Recently, the N-list structure has been proven to be very efficient for mining frequent patterns. This study proposes an N-list-based algorithm for mining FCPs called NAFCP. Two theorems for fast determining FCPs based on the N-list structure are proposed. The N-list structure provides a much more compact representation compared to previously proposed vertical structures, reducing the memory usage and mining time required for mining FCPs. The experimental results show that NAFCP outperforms previous algorithms in terms of runtime and memory usage in most cases.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Tuong Le, Bay Vo,