面向智能电网大数据关联规则挖掘的频繁模式网络模型

Frequent pattern network model for association rule mining of big data in smart grid

针对目前关联规则挖掘频繁树(FP-Tree)算法实现较困难以及难以处理数据库更新的缺点,提出了频繁模式网络(FP-network)模型,将关联规则挖掘所需要的信息压缩到一个无向网络图上,并建立事务项目关联矩阵,从而进行数据存储和数据挖掘。FP-network模型适用于智能电网大数据的关联规则挖掘。以关联规则挖掘在输电线路故障分析领域的应用为例进行算例分析,结果表明所提出的FP-network关联规则挖掘算法不仅继承了FP-Tree算法的优点,而且只需扫描一次数据库,也便于数据库的维护和更新,从而提高了智能电网大数据关联规则挖掘的效率。

Because FP-Tree（ Frequent Pattern-Tree） algorithm for association rule mining is hard to achieve and it also has problems in handling database update, a novel FP-network model is proposed, which innovatively compresses the required data into an undirected network and establishes the transaction-item matrix for the data sto- rage and data mining. The proposed FP-network model is suitable for the association rule mining of big data in smart grid. Case study of the application of association rule mining in power transmission line fault analysis is carried out. The results show that the proposed FP-network algorithm not only inherits the merits of the FP-Tree algorithm, but also only needs to scan the database once, which is convenient for database maintenance and update, and improves the efficiency of association rule mining for big data in smart grid.