基于频繁模式挖掘的实时供应链数据分析

Data analysis in real-time supply chain based on frequent pattern mining

为了从海量的供应链实时数据中发掘隐性的、重要的业务关系,将频繁模式挖掘(FPM)引入到供应链实时数据分析,有利于提高决策的科学性与效率.提出了实时供应链体系结构,包括数据采集层、实时数据处理层与实时供应链应用层,阐述了实时数据采集与处理、频繁路径选择及频繁模式挖掘的方法.详细论述了频繁路径选择、工作流立方建立及频繁模式挖掘3个阶段实时数据处理的内容、原理与方法.结合某服装供应链分销管理,应用DBMiner工具进行了实证研究,结果表明,应用FPM技术进行供应链实时数据分析,可以高效地挖掘与利用实时数据,提高供应链运作的效率与供应链决策水平.

In order to discover the hidden and important business relation, frequent pattern mining （FPM） was introduced into the field of real-time data analysis, which can improve the scientificity and efficiency of supply chain decision. The real-time supply chain framework was proposed, which include data capturing layer, real-time data processing layer, and real-time supply chain application layer. The methods of real- time data capturing and processing, frequent path selection, and frequent pattern mining were discussed. Real-time data processing content, principles and methods in frequent path selection, workflow cube and frequent pattern mining were expounded. A case study of distribution management in some clothes supply chain was conducted with DBMiner tools. Results show that FPM can efficiently mine and exploit real-time data and enhance the efficiency of supply chain operation and the level of supply chain decision.