Maintenance of Discobered Sequential Patterns for Record Modification

Abstract

In the past, some reseatchers proposed efficient incremental mining algorithms for maintenance of sequential patterns as records were insterted or deleted. In addition to record insertion and deletion, record modification is also very commonly seen. Although maintenancee of sequential patterns for record modification can be performed by usage of the deletion procedure and then by the insertion procedure,twice the computation time of a single procedure is needed. It this paper, we thus attempt to design an effective maintenance algorithm for sequential patterns as records are modified. Our proposed algorithm utilizes previously discovered large sequences in the maintenance process, thus greatly reducing numbers of rescanning databasesand improving the performance. Experimental results also show the performance of the proposed approach. It is thus useful for dynamic database mining.