In the real world database updates continuously in several online applications like super market, network monitoring, web administration, stock market etc. Frequent pattern mining is a fundamental and essential area in data mining research. Not only occurrence frequency of a pattern but also occurrence behaviour of a pattern may be treated as important criteria to measure the interestingness of a pattern. A frequent pattern is said to be regular frequent if the occurrence behaviour is less than or equal to the user given regularity threshold. In incremental transactional databases the occurrence frequency and the occurrence behaviour of a pattern changes whenever a small set of new transactions are added to the database. It is undesirable to mine regular frequent patterns from the scratch. Thus proposes a new algorithm called RFPID (Regular Frequent Pattern Mining in Incremental Databases) to mine regular frequent patterns in incremental transactional databases using vertical data format which requires only one database scan. The experimental results show our algorithm is efficient in both memory utilization and execution.
展开▼
