基于紧凑模式树的配电网物资供应链中异常数据挖掘  

Abnormal Data Mining of Distribution Network Material Supply Chain Based on Compact Pattern Tree

作　　者：亓红红 QI Honghong(Zhuhai Power Supply Bureau of Guangdong Power Grid Co.Ltd.,Zhuhai 519000,China)

机构地区：[1]广东电网有限责任公司珠海供电局,广东珠海519000

出　　处：《微型电脑应用》2022年第3期162-164,共3页Microcomputer Applications

摘　　要：配电网物资中的隐含异常数据难以挖掘,导致异常数据挖掘精度不足,提出一种基于紧凑模式树的配电网物资供应链异常数据挖掘方法。通过约束条件把供应链限定在较小的受限数据立方体多维空间内,并计数统计和降序排列构建数据索引表,添加索引编号。根据索引在知识粒度的等价关联、知识、属性分析配电网物资供应链数据,通过粗糙集来链接知识、等价管理分类,描述物资供应链内每一种数据的异常程度。借索引编号顺序组建左斜树,转换异常集添加至左斜树内,形成紧凑模式树向上累积二次挖掘,防止部分数据未被挖掘。仿真实验证明,所提方法能精准识别异常数据,能够提高挖掘精度与挖掘效率。It is difficult to mine the hidden abnormal data in distribution network materials, which leads to the lack of accuracy of abnormal data mining. A mining method is proposed based on compact pattern tree. The supply chain is limited in a small restricted data cube multi-dimensional space by constraints, and the data index table is constructed by counting statistics and descending order, and the index number is added. According to the index, the distribution network material supply chain data are analyzed under the equivalent association, knowledge and attribute of knowledge granularity, and rough set is used to link knowledge and equivalent management classification to describe the abnormal degree of each kind of data in the material supply chain. The left slant tree is constructed by index number sequence, and the transformation exception set is added to the left slant tree to form a compact pattern tree and accumulate secondary mining upward to prevent some data from not being mined. Simulation results show that the proposed method can accurately identify abnormal data and improve the mining accuracy and efficiency.

关 键 词：紧凑模式树 供应链 配电网 异常数据 数据挖掘 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]