关联规则的动态维护及其在财务数据中的应用

关联规则挖掘在许多数据挖掘中有着广泛的应用。当数据库和支持度阈值发生变化时,现有的挖掘方法普遍存在多次扫描数据库或重复遍历复杂数据结构的问题。该文基于增量式更新算法(IUA)和快速更新算法(FUP),提出在数据库与支持度阈值同时变化情况下的关联规则动态维护算法ARDM,并通过Hash结构与模式增长方法进行优化。实验表明:该算法充分利用了已挖掘结果,在数据库和支持度阈值同时变化时比FP-Growth大幅提高了执行效率。最后,将该算法应用于企业财务指标及财务比率分析。

Dynamic maintenance of association rules and its application in the enterprise financial data

Association rules mining is used widely in data mining.However,the cost of repeatedly scanning database and traveling structures is still high when the database is updated and the support thresholds are changed.An association rules dynamic maintenance(ARDM) algorithm was developed using the incremental updating algorithm(IUA) and fast updated algorithm(FUP) to solve the association rules maintenance problem under circumstance of increasing database size and changing support threshold simultaneously.Hash structures and a pattern growth method were introduced to optimize the algorithm.Experiments show that the ARDM performs better than FP-Growth since the algorithm takes full advantage of the mined results.This algorithm is applied to the enterprise financial index and ratio analysis.