Practical Techniques for Speeding Up Enumeration Algorithms for Frequent Itemset Mining Problems

概要

論文の詳細を見る
In this paper, we address practical techniques for efficiently mining all frequent item sets, frequent closed item sets, and maximal frequent item sets, respectively, from transaction databases. We modify the algorithm for enumerating maximal bipartite cliques proposed in SIGAL89 [9, 10], and propose an algorithm for frequent closed item set enumeration. The computation time of the algorithm is theoretically proved to be linear in the number of output, which is not proved for existing algorithms. We further modify the algorithm for practical computation, and for enumerating all frequent item sets, and maximal frequent item sets. By computational experiments, which is done by a competition of frequent set mining algorithms, we showed our approach is quite efficient, especially for realworld problems witch are said to be difficult.
一般社団法人情報処理学会の論文
2004-03-19