挖掘RFID数据库中多维频繁路径的研究

Research on mining multi-dimensional frequent paths in RFID databases

针对应用于物流和供应链管理的射频识别（RFID）系统产生的海量路径数据集中的多维频繁路径挖掘的问题进行了深入的研究，提出了Dim-path与Path－dim两种不同的顺序处理非路径维数据和路径数据的多维频繁路径挖掘算法。这两种算法根据RFID路径数据自身的特点，将RFID数据划分为非路径维数据、位置数据、停留时间数据，将多维路径挖掘问题分解为多维模式分析与序列模式挖掘问题处理，来提高算法的效率。买验结果与算法分析都表明，Dim－path算法与Path－dim算法能够有效快速地挖掘多维频繁路径。

The paper studies deeply the problem of mining the multi-dimensional frequent paths from the gigantic path data set created by a radio frequency identification （RFID） system applied to supply chain management, and proposes the Dim-path algorithm and the Path-dim algorithm, two methods for mining closed multi-dimensional frequent paths in RFID databases. Base on the characteristics of RFID data, the two methods divide RFID data into three parts of the path independent dimensions, the location data, and the duration data, and mine these parts with multi-dimensional analysis and sequential data mining to improve the efficiency of the methods. The experimental and analytical results show that the algorithms of Dim-path and Path-dim can rapidly and efficiently mine the multi-dimensional frequent path.