Fast Updating Maximal Frequent Itemsets Based on Full Merged Sorted FP-Tree

Because of the low efficiency of Maximal Frequent Itemsets(MFI) updating methods, the MFI´s updating methods were analyzed. A new algorithm UAMFI based on Full Merged Sorted FP-Tree (FMSFP-Tree) was proposed. By merging the Sorted FP-Tree and then obtaining the FMSFP-Tree, UAMFI uses the depth-first method to find and update MFI. Finally, the algorithm was tested on the mushroom and T15I4D100K database, and UAMFI´s performances were compared with Mafia. The experimental results indicate that UAMFI is an efficient algorithm for updating Maximal Frequent Itemsets.