An N-list based Parallel Algorithm for Mining Frequent Itemsets

N-list is a novel data structure proposed in recent years. It has been proven to be very efficient for mining frequent itemsets. In this paper, we present PPF, a new parallel algorithm for mining frequent itemsets. The algorithm directly scans dataset to construct a PPC-Tree. Then, the algorithm uses PPC-Tree to generate a series of N-lists which will be assigned to different nodes to mining frequent itemsets. We have conducted extensive experiments to evaluate PPF against PrePost algorithm on four various real datasets. The experimental results show that PPF algorithm is always the fastest one on all datasets.