A EFFECTIVE METHOD FOR FREQUENT ITEMSETS MINING ON VERY LARGE TRANSACTION DATABASE ENVIRONMENT

文件类型：PDF文档
文件大小：896K
PURPOSE : A method for efficiently mining frequent item sets in a very large transaction database environment is provided to rapidly search dense and rare frequent item sets by using a CPFP-Tree. CONSTITUTION : A header table is configured by counting the items of all transactions within an original transaction database, and a CPFP(Compressed Patricia Frequent Patterns)-tree is configured by using the header table. An individual CPFP-Tree for each item is configured by using the CPFP-Tree and the header table(S620). The frequent item sets are searched by using the individual CPFP-Tree(S630).