An Efficient FUSP-Tree Update Algorithm for Deleted Data in Customer Sequences

摘要

In the past, the fast-updated sequential-pattern tree (call FUSP-tree) structure was proposed for mining sequential patterns from a set of customer sequences. An incremental mining algorithm was also designed for handling newly added transactions. Since data may also be deleted in real applications, an FUSP-tree maintenance algorithm for deletion of customer sequences is thus proposed in this paper for reducing the execution time in reconstructing the tree. Experimental results also show that the proposed tree-update algorithm has a good performance than the batch FUSP-tree algorithm for handling the deletion of customer sequences. The proposed tree-update algorithm thus makes the tree update process become easy and efficient.