Parallel computing for mining Frequent Itemset

Frequent Itemset [1] mining is a key step in association rule problem. Early classical algorithms are serial algorithms, such as the Apriori [2], and FP-growth [3]. With the rapid development of information technology, today, the amount of data often needed to handle is based on GB or TB, which has forced the efficiency of mining algorithms significantly improved. At present, more effective method is to use parallel computing to improve efficiency. In this regard, we propose a method based on partition of computing tasks to achieve parallel mining of frequent pattern, and has been experimentally verified in PC cluster.