METHOD FOR GENERATING ASSOCIATION RULES FROM DATA STREAM AND DATA MINING SYSTEM

摘要

A kind of method, for from data flow generate correlation rule extract and data digging system from data flow be provided to effectively production correlation rule extract, be include generate transaction indefinite data add up to. Data digging system adds new node (130) and updates the information of node corresponding with project set according to the appearance frequency and project set of project set comprising in the transaction of generation. The frequent item set that data digging system selection is compared with minimum support ratio combines (135,140), has the supporting rate on project set corresponding with each node of current potential tree.