An Incremental Updating Algorithm for Sequence Patterns Based on Improved PrefixSpan

The PrefixSpan algorithm needs to rerun the algorithm in the process of incremental updating of the sequence pattern, needs to scan the original database repeatedly and needs to construct the large number of projection databases in the mining process. In order to solve these problems, the IPISP algorithm is proposed. The IPISP algorithm uses the sequence pattern base(SPB) to store all the mining sequences and their supports. In the case of incremental mining, the number and size of the projection database are reduced by scanning the sequence pattern base. At the same time, when constructing a projection database, check the prefix of the sequence database with respect to the prefix to avoid constructing the duplicate projection database for the same prefix pattern. The time required to construct the projection database and the size of the projection database are reduced, thereby reducing time and space consumption. Experiments show that the efficiency of the IPISP algorithm is better than PrefixSpan algorithm in the mining of sequential pattern incremental updating.