Algorithm for Frequent Pattern Mining Based on FP-Reference-Tree/List

In order to resolve frequent pattern mining problem efficiently, redundant operation and temporary data in FP-growth algorithm are analyzed, data structure FPR-tree and FPR-list are imported, the method of conditional pattern base generation and storage in FP-growth algorithm are improved, and a new FPRSG algorithm is presented. Theoretical analysis and experiment result both show that FPRSG is more efficient than FP-growth.