An Improved Algorithm for Mining Maximal Frequent Patterns

Mining frequent patterns plays an important role in mining association rules, correlation, multi-dimensional patterns, etc. Since any nonempty subset of a maximal frequent itemsets also is frequent, it is sufficient to mine only the set of maximal frequent itemsets. In this paper, we still base on the FP-tree and propose an optimized algorithm to mine maximal frequent itemsets. By analyzing the data characteristics in FP-tree, before mining further, pruning strategy is used to reduce operations in building corresponding conditional FP-tree. We also present experimental results which show that the performance of our algorithm outperforms the well known FPmax.