Apparatus and method for association rule mining using frequent pattern-tree for incremental data processing

摘要

Disclosed are an apparatus and method for extracting association rules using a frequent pattern tree for processing progressively increasing data. The data sorting unit gradually increases over time and sorts the items corresponding to the plurality of transaction IDs included in the initial transaction data collected for a predetermined time, according to the sorting order in which the incidence decreases. The header table generator generates a header table in which items included in the initial transaction data and the frequentness corresponding to each item are sorted according to the sorting order. The tree generating unit generates a frequent pattern tree composed of nodes including identification codes and frequent degree information of each item included in the initial transaction data. The data updater updates the header table and the frequent pattern tree based on the transaction IDs and items included in the new transaction data sequentially collected over time after the initial transaction data is collected. The frequent pattern extractor searches each node of the frequent pattern tree and sequentially extracts a frequent pattern for each item from an item located at the bottom of the header table. According to the present invention, it is possible to reduce the calculation amount and increase the processing speed for extracting association rules by processing only newly collected transaction data without having to process the entire transaction data in order to extract the association rule from the gradually increasing transaction data. .