Incremental Mining Alogorithm Pre-FP in Association Rules Based on FP-tree

At present, most of association rule mining are static mining algorithms. However, in real Web applications, new transactions are usually incrementally inserted into database, the previous algorithms couldn´t satisfy the rapid growth like this. In this paper, we proposed the incremental mining algorithms called Pre-FP based on PUFP-tree and Per-large-itemsets. This algorithm don´t need rescan the original database when the new transactions surpass the fixed threshold, especially when each time the small fraction of transactions are inserted, Pre-FP can achieve a good execution efficiency for renewing construction of frequent pattern tree.