LCM ver. 2: Efficient Mining Algorithms for Frequent/Closed/Maximal Itemsets 论文

摘要

For a transaction database, a frequent itemset is an itemset included in at least a specied number of transactions. A frequent itemset P is maximal if P is included in no other frequent itemset, and closed if P is included in no other itemset included in the exactly same transactions as P. The problems of nding these frequent itemsets are fundamental in data mining, and from the applications, fast implementations for solving the problems are needed. In this paper, we propose ecien t algorithms LCM (Linear time Closed itemset Miner), LCMfreq and LCMmax for these problems. We show the eciency of our algorithms by computational experiments compared with existing algorithms.