An Efficient Algorithm for Discovering Sequential Patterns in Updating Databases

Abstract

We propose an efficient algorithm to find and update sequential patterns when new transactions are added to an existing database. This method reduces time for scanning the existing and new databases since it uses only transactions that influence the length of sequence. This algorithm outperforms existing algorithm when updated sequential pattern found in the whole database are longer than the patterns in the existing database. Experimental results show the reduction in total execution time.