Progressive Mining of Sequential Patterns Based on Single Constraint
Abstract: Data that were
appeared in the order of time and stored in a sequence database can be processed
to obtain sequential patterns. Sequential pattern mining is the process to
obtain sequential patterns from database. However, large amount of data with a
variety of data type and rapid data growth raise the scalability issue in data
mining process. On the other hand, user needs to analyze data based on specific
organizational needs. Therefore, constraint is used to impose limitation in the
mining process. Constraint in sequential pattern mining can reduce the short
and trivial sequential patterns so that the sequential patterns satisfy user
needs. Progressive mining of sequential patterns, PISA, based on single constraint
utilizes Period of Interest (POI) as predefined time frame set by user in
progressive sequential tree. Single constraint checking in PISA utilizes the
concept of anti monotonic or monotonic constraint. Therefore, the number of
sequential patterns will decrease, the total execution time of mining process
will decrease and as a result, the system scalability will be achieved.
Keywords: sequential pattern
mining, progressive mining of sequential patterns based on single constraint,
progressive sequence tree, big data
Author: Regina Yulia Yasmin
Journal Code: jptkomputergg170171
