Most of the works proposed so far on mining frequent sequences assume that the underlying database is static. However, in real life, the database is modified from time to time. This paper studies the problem of incremental update of frequent sequences when the database changes. We propose two efficient incremental algorithms GSP+ and MFS+. Throught experimetns, we compare the performance of GSP+ and MFS+ with GSP and MFS ― two efficient algorithms for mining frequent sequences. We show that GSP+ and MFS+ effectively reduce the CPU costs of their counterparts with only a small or even negative additional expense on I/O cost.
展开▼