聚类思想在挖掘关联规则中的运用

Cluster Analysis in Mining Association Rules for Huge Transaction Database

数据挖掘中的关联分析技术旨在发现大量数据项集之间有趣的关联关系。虽然A priori算法利用剪枝方法有效地提高运算效率,但在处理超大型事务数据库时,仍会存在内存瓶颈问题。按照项集对数据库进行聚类预处理,然后在各个数据簇内进行关联分析以提高运算效率,且簇的数目可根据情况由数据挖掘者根据情况预先指定。通过对该算法的复杂度分析得出在一定条件下运算的时间复杂度确实有所下降的结论。

Association rule mining finds interesting association relationships among a large set of data items. Although the Apriori algorithm can reduce computing times in its prune step,there is stilt a high complexity in dealing with huge transaction database. Some ideal of cluster analysis are applied in Association Rule Mining to reduce the complexity.