Mining frequent patterns with incremental updating frequent pattern tree

Mining frequent patterns has been studied popularly in data mining research. However, very little work has been done on maintenance of mined frequent patterns. For the real useful frequent patterns, one must continually adjust a minimum support threshold. Expensive and repeated database scans were done. A novel incremental updating frequent pattern tree (IUFP_Tree) structure, which was a dynamic frequent pattern tree for storing compressed information about all frequent patterns, was proposed, and an efficient mining algorithm: IUFP_Miner, for mining the complete frequent patterns was developed. Efficiency of mining was achieved with the following techniques: Database was compressed into a highly condensed data structure; The IUFP_Tree was recycled to avoid repeated database scans; the size of database was gradually reduced by using a trailer table. The performance study shows that the IUFP_Miner method is efficient and scalable for miming frequent patterns, and is an order of magnitude faster than the FP-Growth.