基于FP-growth关联规则的图书馆数据快速挖掘算法研究

Research on Library Data Fast Mining Algorithm Based on FP-growth Association Rules

作为一种模糊关联规则挖掘算法,FP-growth算法在执行效率上明显优于Apriori算法。但是由于模糊属性的不足和空间复杂度较大,导致FP-growth算法在处理大型事务数据库,例如图书馆数据库时,无法实现有效的多层关联规则挖掘。因此,提出一种改进的FP-growth关联规则算法,能够快速向读者进行个性化图书推荐。首先,该算法把大型图书事务数据库根据首项的事务,划分为若干子数据库,并构建相应的子FP-tree结构;然后,采用实时过滤掉层次树中不是频繁项的父项来缩小扫描空间。实验结果表明:相比Apriori算法和标准FP-growth算法,提出的改进FP-growth关联规则算法在运行效率方面有明显提升,为图书的推荐工作提供了科学依据。

As a fuzzy association rule mining algorithm,the FP-growth algorithm is significantly better than the Apriori algorithm in execution efficiency.However,due to the lack of fuzzy attributes and large space complexity,the FP-growth algorithm cannot implement efficient multi-level association rule mining when dealing with large transaction databases,such as library databases.Therefore,an improved FP-growth association rule algorithm is proposed,which can quickly make personalized book recommendation to readers.First,the algorithm divides the large book transaction database into several sub-databases according to the first transaction,and constructs the corresponding sub-FP-tree structure.Then,the scan space is narrowed down by filtering out the parent steps of the hierarchy tree that are not frequent items in real time.The experimental results show that compared with the Apriori algorithm and the standard FP-growth algorithm,the proposed improved FP-growth association rule algorithm has a significant improvement in operational efficiency,which provides a scientific basis for the recommendation work of the book.