UMFP: Update Mining Frequency Paths in RFID System

Radio frequency identification (RFID) technology is fast becoming an important tool for tracking commodities in supply chain management applications. The movement of commodities through the supply chain forms a gigantic workflow which can be mined for the discovery of moving trends that in turn will be valuable in understanding and optimizing business processes. There are some methods dealing with mining frequency paths in static RFID database. But like data stream, RFID system also collects tracking data for moving objects continually. The current mining methods are not created for mining frequency paths in incremental RFID database. In this paper, we propose a method called UMFP to mine frequency paths in incremental RFID database. The method uses existing frequency path mining algorithms to mine frequency paths in each update batch, then merges the results together and returns the frequency paths to users. In our method, the whole RFID data will be mined only once when update batch generated, it will combine the existing data mining results and the last update batch data mining results. The UMFP uses path tree to store mining results. The mining threshold we used in our method is smaller than minimum support threshold which defined by users. By this strategy, UMFP will mine all the potentially frequency paths, ensure that all truly frequency paths will be output when the users request. At last, experiments will show the performances about UMFP.