An Improved Incremental Algorithm for Maintaining Discovered Association Rules
改进的增量式关联规则维护算法

Mining association rules are a major aspect of data mining research, and maintaining discovered association rules are of equal importance. In this thesis, we analyze a previously proposed algorithm IUA and point out its disadvantages and errors and manage to correct these errors. Furthermore we have proposed an improved incremental maintaining algorithm EIUA. Assuming that database \$D\$ is not updated, EIUA has solved the problem of how to maintain discovered association rules efficiently when the two thresholds, minimum support and confidence, change. The experiments have shown the availability and superiority of the new algorithm.