An active candidate set management method using database trigger and incremental update skill for association rule discovery

摘要

PURPOSE: A method for managing an active candidate item of an associative rule probe using a trigger and a gradual update method is provided to automatically probe a mining technique for probing an associative rule using a trigger, a gradual update method, and a database storage procedure in real time. CONSTITUTION: If new transaction(8) is generated, the generated transaction(8) is inserted into a database(6,9,10). Transaction data inserted by an event generation(11) of a gradual candidate item update trigger are extracted and transmitted to a trigger process(13,12). The trigger process(13) combines all combination available candidate items from the transaction data and creates a candidate item set(15,14). The candidate item set(15) is transmitted to a gradual candidate item managing unit(16). The gradual candidate item managing unit(16) classifies the candidate items into an update object(17) and an insertion object(18), and reflects the candidate items to a candidate item set table collectively through a database reflection unit(19). At this time, an update attribute of an update object candidate item is updated by adding '1' to a frequent occurrence coefficient value of tables in a candidate item database.