Frequent itemset mining is one of the fundamental techniques for data mining and knowledge discovery. In the last decade, a number of efficient algorithms have been presented for frequent itemset mining, but most of them focused on only enumerating the itemsets that satisfy the given conditions, and how to store and index the mining result in order to ensure an efficient data analysis is a different matter.
In this paper, we propose a fast algorithm for generating very large-scale all/closed/maximal frequent itemsets using Zero-suppressed BDDs (ZBDDs), a compact graph-based data structure. Our method, "LCM over ZBDDs," is based on one of the most efficient state-of-the-art algorithms proposed thus far. Not only does it enumerate/list the itemsets, but it also generates a compact output data structure on the main memory. The result can be efficiently postprocessed by using algebraic ZBDD operations. The original LCM is known as an output linear time algorithm, but our new method requires a sub-linear time for the number of frequent patterns when the ZBDD-based data compression works well. Our method will greatly accelerate the data mining process and this will leads to a new style of on-memory processing for dealing with knowledge discovery problems.